System and method for generation of user advisories using generative ai in contract analysis

Abstract

A system for contract classification includes a computing device configured to perform several tasks. The device first receives a contract containing multiple contractual elements. It identifies the involved party and the contract's class category. Using an advisory classifier, the system determines a contract parameter based on the contractual elements and calculates an advisory recommendation by considering the parameter, the party, and the class category. The system then ranks the contract according to its parameter and generates a change recommendation. Finally, it notifies the user interface, displaying correlations between specific contractual elements and advisory recommendations.

Description

SYSTEM AND METHOD FOR GENERATION OF USER ADVISORIES USINGGENERATIVE Al IN CONTRACT ANALYSISInventors:Praful SaklaniPedram AbrariAchint AgarwalTITLE OF INVENTIONSYSTEM AND METHOD FOR GENERATION OF USER ADVISORIES USINGGENERATIVE Al IN CONTRACT ANALYSISCROSS-REFERENCE TO RELATED APPLICATIONS rooon This application claims priority to provisional application number 63 / 617505, filed January 4, 2024, provisional application number 63 / 592992, filed October 25, 2023 and provisional application number 63 / 593004, filed October 25, 2023, the entire contents of which are incorporated herein by reference.BACKGROUND

[0002] Human contract analysis, drafting, execution and management has always been plagued by a number of human weaknesses in dealing with complex issues.1. Inconsistency in tracking: The variability in transactions, especially on third-party documents, hinders consistent tracking and collection of trustworthy contract data across different deals.2. Starting over on each new contract: Due to the unavailability of past agreements and summaries, many contracts must be initiated from scratch, wasting time and resources.3. Manual Labor Intensiveness: A lot of time is consumed on tasks such as renewals and negotiations, which are challenging to automate.4. Missed Relationship Opportunities: There's a missed opportunity to use historical contract insights to enhance and personalize customer interactions.5. Lack of Guidance in Drafting: Without insights or recommendations from previous contracts, creating compliant and optimal new contracts becomes problematic.6. Complex Language Barriers: The intricate legal jargon used in contracts confuses business users and frequently necessitates the intervention of specialized legal teams, adding to time and cost overheads.

[0003] Given the inability for humans to have the capacity, a number of platforms have turned to Al to help the humans cope with this complexity in contracts. However, using Al especially Generative Al has resulting in a new and more complex problem set than anticipated.

[0004] Generative Al models lack an inherent grasp of control flows, rules, and logical reasoning, relying instead on statistical patterns from training data. Despite advances inresearch, enabling generative models to have explicit reasoning and rule-following capabilities remains a challenge.

[0005] Techniques such as chain-of-thought prompting only offer a framework for logical sequence of inferences rather than a complete solution.

[0006] While reinforcement learning allows models to adopt basic rule-based behaviors, they still struggle with intricate control flows.

[0007] Predominant generative Al models like GPT can produce fluent outputs but often falter in logical consistency, factual precision, and compliance with intricate rules.

[0008] The reliability of these models diminishes over extensive chains of inference due to the absence of an explicit rule model.

[0009] Alternative approaches using Gen Al focusing on structured knowledge and reasoning, such as semantic parsers and theorem provers, offer more accuracy in modeling complex rules but may not be as fluent or encompassing in knowledge.

[0010] Training a Large Language Model (LLM) to understand contracts, addendums, and associated legal documents is a complex endeavor. One significant challenge is the inherent complexity and ambiguity in legal language. Legal documents often contain phrases that are intentionally broad to cover various scenarios or are couched in terms that have evolved specific meanings over centuries of legal practice. For a model, discerning the nuances and specific intentions behind such phrasings can be difficult.

[0011] Additionally, contracts often reference external laws, regulations, and previously agreed-upon terms. Without a comprehensive contextual understanding, an LLM might misinterpret a clause because it doesn't understand an external reference or how it interacts with the rest of the document. This interconnectedness of legal documents makes them particularly tricky for models that operate primarily on the immediate context they're given.

[0012] Moreover, legal language often evolves based on case law. That is, the interpretation of a phrase in a contract might depend on legal precedents set by courts in prior cases. Without access to this vast body of evolving case law, or without understanding the implications of a particular precedent, an LLM might not provide an accurate interpretation of a contract's terms.

[0013] Cultural and jurisdictional differences add another layer of complexity. Contracts drafted in different countries or legal systems might use similar terms but mean differentthings, based on the local legal tradition and practice. Training a model to understand these distinctions and provide accurate interpretations for diverse legal systems is challenging.

[0014] Lastly, the stakes in interpreting legal documents are incredibly high.Misinterpretations can lead to costly litigation, financial losses, or breaches of regulatory compliance. While LLMs can assist in understanding and drafting legal documents, relying solely on them without human oversight can pose significant risksThe inventors have identified the foul problems with managing Context Windows in LLM

[0015] The first problem identified is a limited ability to process long sequences. LLMs can only process a certain number of tokens at a time, which limits their ability to understand and respond to complex or long-form text.

[0016] Another problem identified is difficulty tracking long-range dependencies. LLMs can struggle to track long-range dependencies in text, which can lead to errors in tasks such as question answering and summarization.

[0017] Another problem identified is increased risk of hallucination. LLMs can hallucinate, or generate false or misleading information, especially when given a large context window.

[0018] Another problem identified is increased computational cost. Training and deploying LLMs with large context windows is computationally expensive.

[0019] Another problem identified is reduced model portability. LLMs with large context windows are less portable and can be difficult to deploy on edge devices.

[0020] Another problem identified is increased risk of overfitting. LLMs with large context windows are more likely to overfit to the training data, which can lead to poor performance on unseen data.

[0021] Another problem identified is difficulty adapting to new domains. LLMs with large context windows can be difficult to adapt to new domains, such as medical or legal text.

[0022] Another problem identified is reduced interp retability. LLMs with large context windows are less interpretable, making it difficult to understand why they make certain predictions.

[0023] A final problem identified is current prompt formats for the LLM often are ineffective at limiting context window errors as they tend to create more ambiguity for the LLM than clarity.FIELD OF THE INVENTION

[0024] The field of the solution relates to computer implemented methods for the automated analysis or use of data, including use of a large language model (LLM), and to related computer implemented methods and systems as used to review and analyze large copra of contracts and other associated documents throughout the life span of the contracts.The Challenge of the LLM Context Window:

[0025] One of the challenges with LLMs is memory management. Given the vastness of language and the complexity of human context, a model can't simply remember everything it has ever seen. Instead, it operates within what's termed a "context window."

[0026] The context window is essentially the immediate span of tokens (words, characters, or subwords) the model considers when generating a response or understanding context. For example, GPT-4, one of OpenAI's LLM variants has a context window of ranging from 2048 32,000 tokens. The context window restricts the model's immediate "lookback." If a piece of information or context appears outside this window, the model won't directly consider it for its immediate response. So, handling long sets of texts is a problem If you feed an LLM a text that exceeds its context window, it will only consider the most recent tokens within that window. For instance, if you input a 3000-token text into a model with a 2048-token window, the earliest 952 tokens won't be directly considered in the model's response. Extending the context window is computationally complex and most LLM models restrict to maintain contextual relevance. Them context window ensures that the model gives more weight to immediate and relevant context. If you're asking the model about a specific topic like termination terms, it will focus on the surrounding information, ensuring that the generated output is pertinent to the current contract or set of contracts.

[0027] Computational efficiency: as mentioned previously, confining the model's focus to a limited number of tokens, managing the context window ensures computational efficiency. Without this restriction, the task of real-time text generation or comprehension might become computationally prohibitive both in cost and size.

[0028] The present solution uses a plurality of methods and processes to overcome this context window issue. An aspect of the solution: Large language models (LLMs) have the potential to revolutionize contract analysis by providing insights that would be difficult or impossible to obtain manually. However, LLMs require a lot of data to train, and they can beinefficient at processing large contracts this solution capitalizes on techniques to improve the use of the LLMs with all their present limitations.

[0029] A variant of the solution allows selectable choices for a user of the system: This solution allows the user to select parts of the contract family, ability to exclude some members. By allowing the user to limit what is sent to the LLM, receives only the documents that are relevant to the task at hand.

[0030] Another variant of the solution manages the context window issue with sophisticated prompts. Prompt language is a way of instructing the LLM to focus on certain aspects of the data. For example, a user might prompt the LLM to identify all the references to a particular clause in a contract. Using prompt variables reduces the amount of data that is passed into the LLM. For example, the user might only prompt the LLM with specific prompts directed at relevant clauses of the contract, rather than the entire document. Another context window management variant of the solution reduces the each contract part to its key provisions. This can be done by identifying the most important clauses in the contract using generative Al and removing the rest of the text.

[0031] Example A contract manager is working on a contract negotiation for a new client. The contract is complex and contains over 100 pages of text. The the manager wants to use an LLM to identify all the references to a particular clause in the contract. The manager first selects the contract family for the client. This includes the contract, as well as all of the other documents that are related to the negotiation, the manager selects only the key documents that are desired, this reduces the number of pages from 300 to 100 The contract manager then uses Generative AT either by structured prompting or other means to reduce the size of the contract to its key provisions. This reduces the size of the contract from 100 pages to 10 pages. Finally, the contract manager prompts the LLM to identify all the references to the particular clause in the contract. The LLM is able to do this quickly and efficiently because it is only processing the key provisions of the contract.

[0032] Another aspect of the solution is it quickly can generate reports based on the user and the outcome desired.BRIEF DESCRIPTION OF THE SEVERAL VIEWS OF THE DRAWINGS

[0033] To easily identify the discussion of any particular element or act, the most significant digit or digits in a reference number refer to the figure number in which that element is first introduced.

[0034] FIG. 1 illustrates an aspect of the hardware components in accordance with a variant of the solution.

[0035] FIG. 2 illustrates an aspect of the ingestion and processing of complex documents in accordance with a variant of the solution.

[0036] FIG. 3 illustrates an aspect of the ingestion and processing of complex documents in accordance with a variant of the solution.

[0037] FIG. 4 illustrates an aspect of the user selectable interface in accordance with a variant of the solution.

[0038] FIG. 5 illustrates an aspect of the Generative Al assistance in accordance with a variant of the solution.

[0039] FIG. 6 illustrates an aspect of a report format in accordance with a variant of the solution.

[0040] FIG. 7 illustrates an aspect of a report and draft agreement in accordance with a variant of the solution.DETAILED DESCRIPTION

[0041] The disclosed technology encompasses systems, methodologies, and / or computer program commodities at varying degrees of technical integration. Such a computer program commodity may comprise a machine-readable storage medium (or multiple mediums) bearing machine-executable instructions to prompt a processor to execute components of the specified technology.

[0042] This machine-readable medium is a physical entity capable of maintaining and storing instructions to be utilized by an instruction execution apparatus. The medium could be, for example, but not restricted to, electronic, magnetic, optical, electromagnetic, semiconductor storage devices, or a fusion of these. A non-limiting list of specific instances of the machine- readable medium includes portable computer diskettes, hard drives, RAM, ROM, EPROM or Flash memory, SRAM, CD-ROMs, DVDs, memory sticks, floppy disks, and mechanicaldevices like punch-cards or tangible structures with instructions. It should be clarified that the aforementioned medium does not consider transitory signals in isolation, like free-propagating electromagnetic waves or electrical signals over wires.

[0043] The machine-executable instructions detailed can be transferred to diverse computational devices from the machine-readable medium or an external computer or storage via networks like the Internet, LANs, WANs, or wireless networks. Such networks may integrate copper or optical fibers, wireless transmission mechanisms, routers, firewalls, switches, gateway computers, and edge servers. Within each computational device, a network interface or adapter fetches the instructions from the network, forwarding them for retention in the device's machine-readable medium.

[0044] Instructions facilitating operations of this technology might be encoded as assembler instructions, ISA instructions, machine codes, microcodes, firmware instructions, circuit configuration data, or code (both source and object) in diverse programming languages. Examples include but aren't restricted to object-oriented languages like Python, Java, C++, and procedural ones like the "C" language. These instructions might operate wholly on a local computer, partly on local and remote computers, or entirely remotely. Remote computers can be linked via networks, inclusive of the Internet via ISPs. In certain cases, hardware such as FPGAs or PLAs could employ the instructions, utilizing their state data to modify the hardware to actualize facets of the technology.

[0045] The technology's facets are expounded with reference to flowcharts and block diagrams of methods, systems, and computer program products per its embodiments. Each block in these can be realized via machine-executable instructions. These instructions could be presented to a processor in general-purpose computers, specialized computers, or other programmable data apparatuses, crafting a machine that institutes the functions denoted in the diagrams. Furthermore, these instructions could be conserved within a machine-readable medium directing device to operate in a specific fashion. The instructions could also be loaded onto a computer or device to prompt a sequence of tasks producing a computer-driven process.

[0046] The depicted flowcharts and diagrams exhibit potential system, method, and product architectures and functionalities per the technology's embodiments. It's essential to note that these blocks, or their combinations, can be realized by hardware systems specifically designed for those tasks or combinations of hardware and machine instructions.

[0047] While multiple embodiments have been detailed, skilled individuals will recognize that modifications can be made without diverging from the broader aspects of the technology. The term "user" here pertains to any individual or entity interacting with the described contract analysis and generation system. "User device" signifies any computational apparatus, from PCs to mobile devices like smartphones, laptops, wearables, and more, employed to access the system. The term "entity" encompasses organizations or user groups engaging with the system, such as businesses or companies either operating or accessing the system.

[0048] This solutions generative Al contract management offers a sophisticated process and system of advanced tools designed to streamline and enhance the handling of business contracts. This solution leverages large language models (LLMs) grounded in comprehensive contract data to provide accurate and context-aware assistance throughout the contract lifecycle.

[0049] Variants of the solution will comprise:1. Contract Summary: Users can employ specific Al prompts to efficiently distill the essence of contracts, orders, and organizational playbooks. The LLM is grounded in a vast corpus of legal documents, ensuring summaries are accurate and contextually relevant.2. Lineage and Timeline Tracing: The solution traces the lineage of contracts, capturing initial agreements and subsequent amendments. This feature is enhanced by grounding the LLM in other historical contract data, allowing it to understand and interpret changes over time.3. Order and Negotiation Summaries: Provides concise summaries of orders and negotiations, eliminating the need for extensive research. The LLM is grounded in both contract- specific data and broader business context to offer insightful summaries.

[0050] Data Grounding for LLMs: The solution implements robust data grounding techniques to ensure the LLM's outputs are reliable and relevant. Grounding data connects the LLM's knowledge to real-world facts and figures, similar to teaching children about concepts through direct experience rather than abstract definitions. This solution variant addresses several challenges:1. Real-life interaction: Grounding helps the LLM understand and use real-world data, making its responses more accurate and reliable in the context of contract management.2. Context relevance: It enhances the LLM's ability to generate contextually relevant answers by connecting abstract legal concepts to real-world examples and specific contract details.3. Error reduction: Grounded LLMs are less likely to produce Al hallucinations, as they're anchored to trusted sources of information, crucial for maintaining accuracy in legal documents.4. Flexibility: LLM grounding allows the Al model to adapt to changes in contract law, company policies, and industry regulations more effectively.

[0051] The solution variants use these various techniques for grounding LLM data.1. Document Embedding: Contract documents are converted into vector embeddings, allowing the LLM to efficiently search and retrieve relevant information.2. Knowledge Graph Integration: A contract- specific knowledge graph is constructed, linking related terms, clauses, and concepts. This helps the LLM understand complex relationships within and between contracts.3. Pre-training on publicly available legal / contract corpora: The LLM is pre-trained on diverse legal datasets to enhance its foundational knowledge of contract law and legal language.4. Fine-tuning on domain- specific data: The LLM is further fine-tuned on anonymized contracts and legal documents specific to the company or industry, enhancing its understanding of specialized terms and practices.5. Retrieval-Augmented Generation (RAG): When generating responses, the system uses RAG to retrieve relevant contract data and incorporate it into the LLM's output, ensuring accuracy and reducing hallucinations. This technique injects the LLM with trusted enterprise data in real-time during response generation.6. Incorporating external knowledge bases: The solution links the LLM to external legal databases and industry-specific knowledge bases for up-to-date information on regulations and best practices.7. Post-processing and validating: The solution conduct post-processing checks to validate LLM outputs against specific legal criteria or thresholds, catching and correcting errors before the final response is delivered.8. Contextual Understanding: By grounding the LLM in company- specific data, industry regulations, and legal precedents, the system provides advice and insights tailored to the user's specific business context.9. Version Control and Diff Analysis: The grounded LLM can analyze different versions of a contract, highlighting changes and potential implications based on its understanding of legal language and business practices.10. Compliance Checking: Leveraging its grounded knowledge of regulations and company policies, the LLM can automatically flag potential compliance issues in contracts.11. Intelligent Clause Suggestion: Based on the context of the contract and grounded data from similar agreements, the solution can suggest relevant clauses or modifications to existing ones.

[0052] By implementing these advanced grounding techniques, the generative Al contract management solution ensures that its outputs are as contextually accurate and up to date as possible. This approach significantly enhances the solution's reliability, reducing the risk of errors or misinterpretations in contract handling, which is crucial in legal and business contexts.

[0053] As grounding techniques continue to advance, future iterations of this system will likely become even more accurate and reliable, potentially expanding its application to more complex legal scenarios and a wider range of industries and domains.

[0054] Next, the Compare Playbook tool harnesses the power of generative Al to assess and compare contracts against the organization's negotiation playbook. This includes a deep examination of both historical and draft agreements to identify deviations. Each deviation is analyzed for its relative risk, ensuring previously negotiated deals' compliance and alerting users to potential pitfalls in draft agreements. By pinpointing deviations and their associated risks, the solution recommends the necessary personnel for approval, ensuring adherence to organizational policies.

[0055] Finally, the Draft Contract feature identifies deviations in executed agreements that surpass organizational risk tolerance. In light of this the draft contract tool can automatically generate draft amendments. This draft retains the core essence of the agreement but incorporates language tweaks to ensure alignment with the company's negotiation standards. By doing so, it drastically reduces the manual workload for legal teams, enabling them toconcentrate on refining the auto-generated language, thereby accelerating the deal cycle time, and ensuring robust compliance with company policies.

[0056] For the purposes of this solution description, the RAG process is as follows: When given a query or prompt, the solution first retrieves relevant information from its contract knowledge base. In this step, the system receives an input, which could be a question, a task description, or any other prompt. Before the language model processes this input, a retrieval solution comes into play. This retrieval system searches through a large database or knowledge base (of general contract data or a specific database), which in the case of contract management, a repository of contracts, legal documents, and regulatory information. The retrieval system uses advanced algorithms (often based on semantic search or vector similarity) to find information that's relevant to the input query. For example, if the query is about force majeure clauses in international contracts, the system might retrieve examples of such clauses, relevant case law, and recent regulatory changes affecting these clauses.

[0057] This retrieved information is then provided to the language model as additional context. As relevant information is retrieved, it's not immediately output to the user. Instead, it's prepared and formatted to be input into the language model alongside the original query. This step is crucial because it's essentially expanding the context that the language model has to work with. In our contract management example, the language model would receive not just the original query about force majeure clauses, but also the retrieved examples, case law, and regulatory information. This additional context is typically appended to the original query or inserted in a structured format that the model has been trained to understand. The GenAI LLM generates its response based on both its pre-trained knowledge and the retrieved information." Now, the GenAI LLM processes both the original query and the additional retrieved information. The LLM isdynamically combining its broad understanding (from pre-training) with specific, up-to-date information from the retrieval step. In generating its response, the model can refer to specific details from the retrieved information, ensuring accuracy and relevance. It can also use its pre-trained capabilities to understand the context, draw inferences, and present the information in a coherent and useful manner. For our force majeure example, the model might generate a response that not only explains what force majeure clauses are (using its pre-trained knowledge) but also cites specific recent examples, refers to relevant case law, and mentions any new regulations that affect how these clauses are interpreted or enforced.

[0058] FIG. 1 provides just a variant of the solution of a computer system. The solution comprises a User Device 102, a Network 104, a SaaS Contract Platform 108, a Data Repository 110, an Al Training Module 612, a Training Data Repository 114, an Ingested Contracts 116, an Advisory Recommendation 118, and a Contract Draft Generator 120. As displayed in the FIG. l, one or more user devices are operationally linked through a network to one or more entity systems. Consequently, the user devices can transmit and retrieve data from the entity systems. In the depicted embodiment, several user devices might connect with the entity systems via the network. In alternative configurations, certain systems, devices, or servers might be amalgamated into a single unit, or be divided into multiple units. It should be inferred that the various servers, systems, and devices can be amalgamated in different variants while retaining analogous functionality.

[0059] For example, the network might be a dedicated distributive network which identifies specific network triggers. Moreover, the network could be a Global Area Network (GAN) like the Internet, a Wide Area Network (WAN), a Local Area Network (LAN), or any fusion of such networks. This network can facilitate wired, wireless, or combined communications. Additionally, the network might encompass a peer-to-peer communication system.

[0060] Both the entity systems and user devices possess data storage and processing capabilities, holding data pertinent to the system environment, including data devised by various system applications. The contract analysis and generation system's diverse applications and their associated data may reside within the memory of the entity systems and user devices. In some versions, users might engage with the contract analysis and generation system via a user interface. This interface is designed for simple contract generation.

[0061] Numerous modules and repositories are also detailed. The user interface is equipped for single-click contract analysis. A particular data repository various contract clauses and verifies the inclusion of all vital clauses in a contract. Another module employs an Artificial Intelligence (Al) model trained on queries to ascertain if humans input questions possess default responses and is engineered to insert these answers with contextual output in an answer to the questions.

[0062] Another Al module comprises keywords and links to various regulations (like Federal Acquisition Regulations with results tuned to export the unique and necessary contract wording.

[0063] Another Al model is educated to extract contract terms from a list of questions. A distinct repository is designed to exclude keywords identified by Al Named-Entity Recognition (NER) that aren't viewed as variables in a contract from the working text.

[0064] Additional Al models are educated on keywords and questions, converting keywordcontaining sentences into ordinarily written questions.

[0065] Yet another Al model classifies anticipated contract issues into different issue classes.

[0066] The machine learning module conducts various comparisons like comparing templates, contracts against templates, and distinct contract versions This module can also analyze party details into question lists and contract templates. Furthermore, it might be linked to both memory and the user interface. The contract lifecycle management module is engineered to oversee a contract's lifecycle.

[0067] In some versions, users and entities might utilize ERP and CRM systems as data repositories. These systems might retain data about users, entities, and their devices. Certain modules might be accessible by the party management module using established Application Programming Interface (API) techniques. ERP and CRM systems might reside on the entity systems or user devices.

[0068] To address the context window limitation inherent in Large Language Models (LLMs) like GPT-3 or GPT-4, a multi-phase prompting methodology is used in this solution. As an example, below a 2-phase approach is disclosed. This approach involves the creation of preliminary prompts to distill relevant information using a Generative Al approach, followed by the aggregation of these prompts into a set of final, contextually rich input prompts. This method is designed to circumvent the token limit constraints of the LLMs' context window, ensuring a more accurate and coherent model response.Phase 1: Preliminary Prompt Engineering

[0069] Information Extraction via Selective Querying: Initial prompts are designed to elicit specific, targeted responses. This involves a careful selection of queries to extract only the most pertinent information required.

[0070] Token Economy: Given the token budget [as well as the cost of the prompt] constrained by the model's context window, each preliminary prompt and its response are optimized for brevity while maintaining informational integrity.

[0071] Data Aggregation and Truncation: Responses to these preliminary prompts are then aggregated. This step often involves truncating or summarizing the extracted data to fit within the token limitations, ensuring that only the most relevant context is retained.Phase 2: Synthesis into Final Prompts

[0072] Contextual Concatenation: The refined data set from Phase 1 is then concatenated into a set of final prompts. Each final prompt is carefully engineered to include all necessary context within the LLM's token limitations.

[0073] Optimization for Contextual Coherence: Each final prompt is optimized for coherence and relevance. This involves strategic structuring of the prompt to place the most critical information within the immediate observation window of the LLM, ensuring maximum retention of context during the model's inference process.Technical Advantages of This Methodology

[0074] Enhanced LLM Model Performance: By providing a distilled yet comprehensive context, the model can generate responses with higher accuracy and relevance, mitigating the risk of context dilution or loss.

[0075] Efficient Utilization of Computational Resources: This approach minimizes the computational load and cost by avoiding redundant or irrelevant data processing, aligning with efficient resource utilization principles in computational linguistics.

[0076] Adaptive Context Management: The methodology provides a flexible framework to adapt the information density and context breadth based on the specific requirements of the task, offering a customizable solution to context window limitations.

[0077] Information Density Optimization: By maximizing the information density within the token limit, the approach ensures that the model operates with the richest possible context, enhancing the quality of language generation and understanding.

[0078] In FIG. 2 Illustrates a configurable multi-phase prompt flow of a compound document set (e.g., A master contract, addendum, an order, etc.). The prompting flow comprises a Data Repository 202, a Solution control 204, a Preliminary Prompt 1 206, a Preliminary Prompt 2 208, a Final Prompt 1 210, a Final Prompt 2 212, a GenAI LLM 216, a Micro Call 1 218, a Micro Call 2220, a 226, and a Data Dependency 228. Taking the compound document from the Data Repository 202, and then entering the document into Solution control 204. Solutioncontrol 204 breaks up the document into selected segments of text first into Preliminary Prompt 1 206 which sends selected segments of text to the GenAI LLM 216 in a series of Micro Call 1 218, Micro Call 2 220, to Micro Call X . The Solution control 204 then repeats this by sending the next preliminary set of document segments to the GenAI LLM 216 in a similar Micro Call 1 218, Micro Call 2 220, onward to a Micro Call X. Each of the preliminary prompts are chosen by Solution control 204 to get an optimized result for analysis of a compound document.[00791 For the sake of clarity there can be multiple Data Repository 202 of different types including knowledge graphs, RAG enhanced searching or vector databases. These can be accessed sequentially or in parallel.

[0080] An example below is described using a contract as an example.

[0081] Preliminary Prompts These prompts are used to preprocess contracts data into a form that can be used by the final prompt(s) to provide the final solution output. These prompts must have prompt variables (e.g. { {MASTER} }, { {ORDER} }, ...) that dictate which contracts data is passed into the prompt. The output of these preliminary prompts are streamed to the screen and ultimately passed to final prompt. They are also reusable across compound prompts.

[0082] Document Processing Configuration There are two configurations: ALL, ONEBYONE. ALL means all instances are passed in together. ONEBYONE means the docs are passed to LLM one-at-a-time in order of effective date, as shown:

[0083] Doc Family contracts processed in order of effective date.• If prompt refers to both { {MASTER} } and { {AMENDMENT} } then only the AMENDMENTS that have a MASTER as a parent are processed in order of effective date• If prompt refers to both { {ORDER} } and { {AMENDMENT} } then only the AMENDMENTS that have an ORDER as a parent are processed in order of effective date• If prompt refers { {AMENDMENT} } without { {MASTER} } or { {ORDER} } then all AMENDMENTS are processed in order of effective date.

[0084] Final Prompt(s) The final prompt(s) are different than the preliminary prompts in that they always get the output of the preliminary prompts in their CONTEXT (via HISTORYASCONTEXT). If the prompt has other prompt variables, they corresponding docs are also passed in as part of the CONTEXT section. There is no one-at-a-time processing option for final prompt(s). The output of final prompts is streamed to the screen. InputConfiguration Each final prompt can be configured to have one or more of the following inputs passed in as part of its CONTEXT (along with its prompt variables):• Output of 0 or more preliminary prompts. ALL configuration would pull in all.• Output of 0 or more final prompts preceding it. ALL configuration would pull in all.

[0085] Each preliminary prompt must be referenced by at least one final prompt in the compound prompt.

[0086] Prompt Input Context: The input context of a prompt is processed based on processing pattern. Final prompts must include HISTORYASCONTEXT in the pattern.

[0087] Document Generation Configuration Final prompts can optionally generate a Word document. If this option is configured, then a corresponding template must be selected from the Memos folder under Templates. Prompt variables after the Referenced Documents section can be mapped to preliminary prompts and final prompts. But each variable can only be mapped to one and only one prompt.

[0088] An example of Compound Prompting in this solution is presented below.

[0089] Compound Prompt Configuration Each compound prompt is configured as a series of optional preliminary prompts followed by one or more final prompt(s).

[0090] Degenerate compound prompts (No preprocessing prompts, 1 final prompt):• No document generation: o Summarize Master o Summarize Orders o View Playbook• Document generation: o Draft New Amendment o Compliance Review (Latest)• Compound prompts (1 or more preprocessing prompts, 1 or more final prompts): o No document generation:■ Compliance Review (Historical):• Document generation: o Account Review Memo: o Preliminary Prompts :■ Summarize Master■ Summarize Orders o Final Prompt■ Negotiation Guidance• Draft Contract Redlining & Negotiation o Preliminary Prompts:■ Compare Draft Contract to Previous o Final Prompt■ Negotiation Guidance and Redlines

[0091] Doc Family - Prompt Input Configuration There is a prompt level configuration to determine how the data from doc family contracts is passed into the prompt:• Contract Metadata• Contract Full Text• Both

[0092] In FIG. 3 an example of how this compound prompt and flow control would work in an account review of a contract, orders and extraction of negotiation guidance The system comprises the components of Data Repository 302, a Solution Control Flow 304, a Summarize Master 306, a Summarize Order 308, a Negotiation Guidance 310, a Solution output 312, a GenAI LLM 316, a Micro Call 1 318, a Micro Call 2 320, a 326, and a Data Dependency 328. After selecting an account for review from the Data Repository 302, the Solution Control Flow 304 performs the following processes.

[0093] Curated Text Input:• Selection: Choose selection of Master Agreement and its subordinate documents text segments, considering relevance, quality, and potential biases.• Preliminary Preparation: Format text appropriately in a preliminary prompt format [e.g., Summarize Master 306, Summarize Order 308 to be sent into the GenAI LLM 316 processing (e.g., tokenization, cleaning, encoding).

[0094] 2. Control Flow in LLM:• Preliminary Prompt: Compose a prompt that effectively guides the LLM to: o Understand the context and goals of the task. (Summarize Master 306) o Utilize the curated text appropriately.• Task Execution: The LLM processes the prompt and curated text in a sequential manner:o Lexical Analysis: Breaks down text into words and tokens. o Sentence Parsing: Analyzes grammatical structure and relationships. o Semantic Interpretation: Extracts meaning and context. o Knowledge Integration: Connects information from curated text with its internal knowledge base. o Response Generation: Synthesizes a response based on analysis and knowledge.

[0095] 3. Preliminary Result:• Review and Evaluation: Analyze the initial output for accuracy, relevance, and alignment with goals.• Decision-Making: Choose whether to proceed with cascading or refine the prompt or input.

[0096] 4. Cascading Prompt [Negotiation Guidance 310]:• Construction: Formulate a new prompt that: o Incorporates insights from the preliminary result. o Incorporates o Guides the LLM towards a more accurate or refined response. o Potential enhancement techniques for iteration:■ Highlighting specific aspects of the curated text.■ Reframing the task with additional context or constraints.■ Addressing identified errors or shortcomings.• Re-Execution: The LLM processes the cascading prompt, leveraging previous knowledge and analysis to produce a more refined response.

[0097] 5. Solution Output:• Solution output 312 is then what is expressed after the Negotiation Guidance 310 is returned from the Gen Al LLM 216.• Evaluation and Refinement: Assess the final output against desired criteria.• Iterate as Needed: Repeat the process with further cascading prompts or input adjustments until a satisfactory response is achieved.

[0098] Example of the Solution output 312

[0099] A Master / Order Reference TableEXAMPLE OF NEGOTIATION GUIDANCE

[0100] Based on the [Customer Playbook] and the current terms of the Master Agreement, Amendments, and Orders, the following changes are proposed:TERM & RENEWAL• Current Provision: The initial term of the agreement is 24 months with automatic renewal for additional terms of 12 months unless either party provides notice of its intention not to renew, in whole or in part, by providing written notice at least 60 days prior to the expiration of the Initial Term or any Renewal Term. [FIRST AMENDMENT TO THE MASTER AGREEMENT | 10883 | 1]• Proposed Change: “This Agreement shall commence on the effective date and continue for 2 years, automatically renewing for subsequent one-year periods unless either party provides written notice of non-renewal at least 30 days before the end of the current term.”ASSIGNMENT• Current Provision: Neither Party will assign, subcontract, delegate, or otherwise transfer this Agreement, or its rights and obligations herein, without notifying the other Party of their intent at least 90 days in advance. [SECOND AMENDMENT TO THE MASTER AGREEMENT | 10884 | 1]• Proposed Change: “Vendor may assign this Agreement, in whole or in part, without the prior written consent of Customer. Customer may not assign this Agreement without the prior written consent of Vendor.”PAYMENT TERM• Current Provision: All Charges are due net 60 days from the receipt of invoice;[Client] will invoice Client monthly; and all payments made under this Agreement will be in US Dollars. [Master Services Agreement | 10881 | 5.1]• Proposed Change: “Customer will pay all invoiced amounts within 30 days of the invoice date. Late payments may incur a charge of 3% per month on the unpaid balance.”TERMINATION• Current Provision: If either Party materially breaches the Agreement, the nonbreaching Party may provide a written notice of termination for cause. The breaching Party shall have thirty (30) calendar days from receipt of such notice to cure the breach, provided that such breach is curable. [Master Services Agreement | 10881 | 6.3]• Proposed Change: “Either party may terminate this Agreement for any reason upon 60 days’ written notice to the other party. Additionally, either party may terminate immediately in the event of a material breach by the other party.”PRICE INCREASE• Current Provision: [Client] may change the Charges for [Client] Solutions upon sixty (60) days’ notice to Customer no more than once per year and by no more than 3%. [Master Services Agreement | 10881 | 5.2]

[0101] Proposed Change: “The fees under this Agreement may be increased each year by no more than 5% of the fees in the preceding year.”MASTER SUMMARY EXAMPLE[ Master Services Agreement | 10881 | Master Agreement]• Term, Renewal, Expiration Dates• Effective Date'. June 12, 2019• End Date'. June 11, 2023• Status'. Active• Summary: The agreement provides for an initial term with auto-renewal, payment terms of net 60 days, and termination for cause with a 30-day notice period. Price changes are allowed once per year with a 60-day notice and arc capped at a 3% increase.• Key Provisions: o Term & Renewal: This MSA will be in effect for 12 months from the Effective Date of this MSA. CloudPro may terminate the Agreement and any Order or Statement of Work in whole or in part upon six (6) months’ written notice to the Customer in relation to a CloudPro Solution which is being discontinued by CloudPro.■ Term Type: Initial Term with Auto-Renewal■ Renewal Type: Auto-Renewal■ Renewal Term (months): 12■ References: [Master Services Agreement | 10881 | 6.1]o Assignment: Neither Party will assign, subcontract, delegate, or otherwise transfer this Agreement, or its rights and obligations herein under any circumstance.■ References'. [Master Services Agreement | 10881 | 15] o Payment Term: All Charges are due net 60 days from the receipt of invoice; CloudPro will invoice Client monthly; and all payments made under this Agreement will be in US Dollars.■ Payment Due'. 60 days■ References'. [Master Services Agreement | 10881 | 5.1] o Termination: If either Party materially breaches the Agreement, the nonbreaching Party may provide a written notice of termination for cause. The breaching Party shall have thirty (30) calendar days from receipt of such notice to cure the breach, provided that such breach is curable. If the breach, is not cured within such period, the Agreement will terminate as indicated in the notice.■ Termination Type'. Termination for Cause■ Termination Notice Period'. 30 days written notice.■ References'. [Master Services Agreement | 10881 | 6.3] o Price Increase: CloudPro may change the Charges for CloudPro Solutions upon sixty (60) days’ notice to Customer no more than once per year and by no more than 3%.■ Price Lock Period'. 1 year■ Price Increase Frequency. Once per year with a 60-day notice■ Price Increase Caps'. 3%■ References'. [Master Services Agreement | 10881 | 5.2] Orders SummaryMASTER SUMMARY EXAMPLE[ [Product] Quote and Order | 10885 | Order]• Term, Renewal, Expiration Dates• Effective Date'. August 1, 2019• Governing Terms'. Master Services Agreement effective date June 12, 2019 (ID Number: 10881)• Order of Precedence'. In the event of conflict between the Terms of this Order and the Master Services Agreement, this Order shall take precedence.• End Date'. July 31, 2023• Status: Active• Summary: The contract stipulates an initial term of 12 months with automatic renewal for additional terms of one year each, unless a notice is provided at least 45 days prior to the end of the initial term or any renewal term. Payments are due 30 days after receipt of an invoice. Price increases are allowed upon 60 days’ notice prior to the end of the initial term or any renewal term but cannot exceed 2%.• Key Provisions: o Term & Renewal: Customer’s Subscriptions shall have an Initial Term of 12 months commencing on the Effective Date. Following the end of the Initial Term, the Subscription will automatically renew for additional terms of [one year], each a “Renewal Term” unless Customer provides notice to CloudPro at least 45 days prior to the end of the Initial Term, or any Renewal Term.■ Term Type: Initial Term with Auto-Renewal■ Renewal Type: Auto-Renewal■ Renewal Term (months): 12■ References: [[Product] Quote and Order | 10885 | Initial Term, Renewals] o Payment Term: Unless otherwise specified on the applicable invoice, payment is due 30 days after Customer’s receipt an invoice. Credit cards are not accepted for payment. Customer is responsible to pay any applicable taxes related to the purchase of its Solution.■ Payment Due: 30 days■ References: [[Product] Quote and Order | 10885 | Payment] o Assignment: Not specified o Termination: Not specified o Price Increase: CloudPro may increase the Customer’s Monthly Subscription Fees upon 60 days’ notice prior to the end of the Initial Term, or any Renewal Term by no more than 2%.■ Price Lock Period: 1 year■ Price Increase Frequency : Every year with a 60-day noticePrice Increase Caps'. No more than 2%

[0102] FIG. 4 is an example of a playbook generated by solution variant. A unique feature of this solution will generate a playbook as a set of advisory points or guidance for a less experienced negotiator to prevent the user from giving away negotiated points that open the organization to risk. For this specification, the term “Advisory” is intended to be construed as a superset of guidance element which comprise; risk advisory, content advisory or timeline advisory given as variants of the solution. Examples of Risk Points and Guidelines beyond the exemplar drawing would also comprise the following:

[0103] Solution Risk Point Identification:1. Flagged Unique Contract Type: Without clarity on the type of the contract, (e.g a services contract format used for product delivery.)2. Undefined or Ambiguous Parties: Often 3rd party obligations creep into contract language, by not clearly identifying all involved parties can lead to future disputes.3. Unclear or Conflicted Duration: Without a well-defined duration and renewal terms, there may be unintended contract extensions or premature terminations.4. Conflicting Payment Terms: Ambiguous payment conditions can result in financial discrepancies.5. Termination Ambiguities: Unclear termination conditions might lead to unintended penalties or breach of contract.6. Loose or Missing Confidentiality Clauses: Inadequate confidentiality or non-disclosure terms can risk leakage of sensitive information.7. Governing Law: Agreeing to jurisdictions that are hostile to the organization is a significant risk point.8. Undefined Dispute Mechanisms: Without clear dispute resolution methods, disagreements can become protracted or costly.9. Overlooked or Unique Special Provisions: Missing out on unique or unusual terms can introduce unexpected obligations or rights.10. Unchecked Liabilities: Not assessing potential liabilities can expose the organization to undue risks.

[0104] Solutions Categories for Guidelines for Procurement1. Clearly Define the Contract Type: Limiting the types of contracts to be accepted.2. Identify All Parties and update negotiation playbook: Ensure all entities involved in the contract are clearly listed.3. Establish Clear Contract Duration: Limit contracts to the minimums and maximums.4. Detail Payment Terms: Lists of cleared payment amounts, frequencies, methods, and conditions.5. Clear Termination Clauses: List the minimum and maximum grounds for contract termination and any associated penalties.6. Confidentiality Minimum Guidelines: Ensure robust clauses for confidentiality or nondisclosure are in place.7. Governing Law as listed: Clearly state the legal jurisdiction governing the contract.8. Permitted Dispute Resolution Mechanisms9. Flagging Special Provisions:10. Flagging other solution Risks and Liabilities as needed

[0105] FIG. 5 is an example of a historical look back at previous contracts with an entity or entity group that would inform the negotiations of a current contract. Other aspects of historical risk extracted from historical contracts also comprise:1. Problem Clauses: Previous contracts might contain clauses that have historically resulted in disputes or unfavorable outcomes.2. Performance on Payment: Past contracts may have ambiguous or inconsistent payment terms and performance that caused financial discrepancies or disputes.3. Problem or Missing Definitions: Definitions that have caused historical disputes.4. Risky Termination Clauses: Historical termination clauses were either too rigid or too lenient, affecting the parties involved in negative formats.5. Mismatched Expectations: Previous contracts might have set expectations that were not met, leading to dissatisfaction or breaches.6. Overlapping Jurisdictions: Historical contracts may have had unclear or conflicting jurisdiction clauses, complicating dispute resolutions.7. Problematic Obligations: Past contracts might not have clearly delineated the responsibilities of each party.8. Inadequate Dispute Resolution Mechanisms: Previous agreements may have lacked a clear mechanism for handling disputes, leading to prolonged or unresolved disagreements.9. Unfavorable Auto Renewal Terms: Past contracts might have auto-renewal terms that were not in a party's favor or went unnoticed.10. Missed Deadlines or Milestones: Historical instances where critical deadlines or milestones were missed, leading to penalties or breaches.11. Weak Indemnity Clauses: Historical contracts may have had inadequate indemnity clauses, leading to financial burdens.[01061 FIG. 6 is an example of Guidelines for a new draft contract based on the solution's assessment. This example recommends three areas of concern for the negotiators.

[0107] Summarize Contracts prompt - this prompt will create a summary of the lineage of contracts for a business relationship, including the original agreement and any follow-on amendments .

[0108] Summarize Orders prompt - summarizes all the orders executed for a business relationship like products sold / purchased renewal dates, etc.

[0109] Summarize Playbook prompt - summarizes the organization’s negotiation playbook converted from raw form into defined rules by Al.

[0110] FIG. 7 shows the construction of a proposed amendment or draft from the playbooks or summaries, that the solution can generate using an LLM to generate from its knowledge base. The Al contract management solution leverages its Large Language Model (LLM) to construct proposed amendments or draft contracts by drawing from a rich knowledge base of playbooks, summaries, and historical data. The solution variant begins with the LLM analyzing the current contract or amendment request, identifying key elements such as parties involved, contract type, and specific clauses requiring modification. The system then delves into the organization's contract playbooks, knowledge graphs or vector databases which contain standardized language, preferred clauses, and best practices. Simultaneously, it searches through corpora of historical contracts and amendments, identifying similar cases or precedents relevant to the current situation. The solution then interprets the required changes, considering factors such as organization or industry-specific terminology, regulatory requirements, and the organization's past negotiation patterns. Armed with this comprehensive understanding, the solution variant generates a draft amendment or contract that incorporates appropriate language from the playbooks, adapts relevant clauses from similar past contracts, and creates updated content as needed. The generated draft undergoes an automatic compliance check against internal policies, regulatory requirements, and industry standards. If amending an existingcontract, the solution variant performs a detailed comparison between the original and the proposed amendment, highlighting changes and potential impacts. To provide transparency and aid decision-making, the solution variant produces an explanation of its choices, referencing specific playbook items, historical precedents, or regulatory considerations that influenced the draft. The solution variant may even generate multiple versions or suggest alternatives for specific clauses, allowing legal professionals to choose the most appropriate option. As the organization's legal experts review and modify the generated drafts, the system learns from these interactions, continuously refining its understanding of preferred language and structuring for future generations. This solution variant not only streamlines the contract amendment and drafting process but also ensures consistency with industry or organizational standards while adapting to the specific needs of each unique contractual situation, serving as a powerful tool that enhances efficiency and reduces the risk of oversight in contract management.

[0111] In conclusion, a generative Al (Gen Al) contract platform can complete tasks a human cannot perform, an example is provided below throughout the lifecycle.

[0112] Contract Lifecycle: Request Stage• Traditional: Contract initiation often happens through emails, without any systematic tracking or visibility, leading to inefficiencies in resource allocation.• Gen Al: Offers streamlined contract initiation by combining conditional logic forms and Al insights, optimizing the process, and providing proactive notifications and risk assessments before the contracts are drafted.

[0113] Contract Lifecycle: Authoring & Approval• Traditional: contract managers often face the challenge of choosing the right template due to a multitude of them, leading to manual effort and inconsistencies.• Gen Al: Leverages Al to automatically generate compliant on its first drafts of contracts, eliminating the overhead of multiple templates.

[0114] Contract Lifecycle: Negotiation• Traditional: Determining what's been signed previously is a manual and tedious process due to a fragmented history of contracts and generally a fragmented set of contract managers performing the negotiations.• Gen Al: Provides on-demand summaries of past negotiations, enabling better decisionmaking and leveraging a collective knowledge unattainable by a human contract group.

[0115] Contract Lifecycle: Storage & Analytics• Traditional: Contracts are often dispersed across different storage systems, leading to data fragmentation and difficulty in accessing required contracts.• Gen Al: Enables a holistic approach to a contract data lake, with automated organization and extraction of key contract data, simplifying storage and retrieval processes.

[0116] Contract Lifecycle: Compliance & Obligation Management• Traditional: After signing a contract, compliance events are manually tracked, leading to missed obligations and potential penalties.• Gen Al: Continually scans past contracts for specific commitments requiring future action, ensuring no compliance event is overlooked.

[0117] Contract Lifecycle: Renewal & Termination• Traditional: Renewal drafts are created manually from scratch, leading to inefficiencies. Termination processes lack systematic reminders, resulting in missed opportunities to exit contracts.• Gen Al: Facilitates the automated creation of renewal drafts based on relationship history and trends. For terminations, it proactively notifies stakeholders and auto-generates termination notices, streamlining the process.

[0118] In conclusion, while traditional contract management practices are often manual and fragmented, Gen Al contract management and curation offers a more streamlined, automated, and insightful approach, leveraging the power of Al to address the challenges and inefficiencies of the traditional method.ADDITIONAL CONSIDERATIONS

[0119] While the preceding description outlines various solution variants in detail, it's important to recognize that the legal scope of the invention is ultimately defined by the claims listed at the end of this patent. The detailed description serves as an illustrative example, and it is not exhaustive of all potential solution variants. Due to the impracticality of describing every conceivable solution variant, alternate configurations may exist — whether using current technologies or those developed after this patent’s filing — that still fall within the scope of the claims.

[0120] Throughout this specification, references to singular instances may also include plural instances, and vice versa. Likewise, while operations of methods are described separately, they can be performed concurrently or in a different sequence than presented. Components orfunctionalities described as separate in example configurations may be combined, while those presented as a single entity may be divided into multiple components. These and other modifications or improvements remain within the bounds of the described invention.

[0121] In certain solution variants, logic, routines, subroutines, applications, or instructions may be executed via software (e.g., code on a non-transitory, machine-readable medium) or hardware (e.g., special-purpose processors). In a hardware context, these routines can be physical, tangible units configured in specific ways, such as through a field-programmable gate array (FPGA) or an application-specific integrated circuit (ASIC). Alternatively, they may leverage general-purpose processors configured temporarily via software to execute specific operations. Decisions on whether to implement routines in dedicated hardware, software, or hybrid solutions may depend on cost, complexity, or other constraints.

[0122] For purposes of clarity, "hardware module" should be understood to mean a tangible entity that can either be physically constructed or configured (permanently or temporarily) to operate in a specific manner. If temporarily configured via software, a general-purpose processor may act as various hardware modules at different times. This flexibility enables the same processor to perform multiple functions dynamically, depending on the system’s current needs.

[0123] Inter-module communication between hardware modules may occur through signal transmission over circuits or buses. When modules are instantiated at separate times, data can be stored and retrieved from shared memory structures, enabling asynchronous operation. For instance, a hardware module may execute an operation and store its results in memory, allowing another module to access and process the stored information later.

[0124] The operations of methods described in various solution variants may be partially or fully implemented by one or more processors. These processors may be physically located within a single machine or distributed across multiple systems, enabling distributed processing. In some cases, these systems may be in a centralized location, like a data center, while in other cases, they could be spread across multiple geographic locations. When processors are distributed, they may communicate and coordinate their tasks via networked infrastructure, forming a cohesive processing system.

[0125] Terminology used herein, such as "processing," "computing," or "calculating," refers to the manipulation of data in physical forms, such as electrical, magnetic, or optical quantities. When the specification refers to "one solution variant" or "a solution variant," it indicates thatthe described feature may be applicable to at least one possible solution variant. This should not imply that all instances of the phrase refer to the same solution variant.

[0126] Additionally, terms like "comprises," "including," and their variants are intended to imply non-exclusive inclusion. For instance, a method that "comprises" certain elements is not limited to those elements alone and may include other components not explicitly listed.Similarly, "or" should be interpreted as inclusive unless otherwise specified, meaning A or B could be true individually or simultaneously.

[0127] The descriptions provided are intended as illustrative, non-exhaustive examples. They do not define every possible solution variant, as doing so would be impractical, if not impossible. Moreover, technological advancements and alternate configurations may arise that still fall within the invention’s defined scope.

Claims

CLAIMSWhat is claimed is:

1. A system for classification of contracts, the system comprising a computing device, wherein the computing device is configured to: receive a contract comprising a plurality of contractual elements; identify a party to the contract and a class category of the contract; determine at least a contract parameter using an advisory classifier as a function of the plurality of contractual elements; calculate an advisory recommendation as a function of the at least a contract parameter, a party, and the class category; rank the contract in terms of its contract parameter; and notify a change recommendation for the contract to a user interface, with at least a correlation between a contractual element and an advisory recommendation.

2. The system for classification of claim 1 creating a reduced size semantic equivalent contractual element of each of the contractual elements in the plurality of contractual elements.

3. The system for classification of claim 1 further comprising a user selectable interface to select a subset of the contractual elements for analysis.

4. The system for classification of claim 1 further comprising a set of predefined prompts that are user selectable to assist an LLM to selectively review a subset of the contractual elements.

5. An apparatus for contract classification comprising: a processor; and a memory storing instructions that, when executed by the processor, configure the apparatus to: receive a contract comprising a plurality of contractual elements; identify a party to the contract and a class category of the contract; determine at least one contract parameter using an advisory classifier based on the plurality of contractual elements;calculate an advisory recommendation based on the at least one contract parameter, the party, and the class category; rank the contract based on the at least one contract parameter; and generate, on a user interface, a change recommendation for the contract including at least a correlation between a contractual element and the advisory recommendation.

6. The apparatus of claim 5, wherein the memory stores further instructions that, when executed by the processor, configure the apparatus to: create a reduced size semantic equivalent contractual element for each of the contractual elements in the plurality of contractual elements.

7. The apparatus of claim 5, wherein the memory stores further instructions that, when executed by the processor, configure the apparatus to: provide a user selectable interface configured to enable selection of a subset of the contractual elements for analysis.

8. The apparatus of claim 5, wherein the memory stores further instructions that, when executed by the processor, configure the apparatus to: provide a set of predefined prompts that are user selectable to assist a Large Language Model (LLM) to selectively review a subset of the contractual elements.

9. A method for contract classification comprising: receiving, by a computing device, a contract comprising a plurality of contractual elements; identifying, by the computing device, a party to the contract and a class category of the contract; determining, by the computing device, at least one contract parameter using an advisory classifier based on the plurality of contractual elements; calculating, by the computing device, an advisory recommendation based on the at least one contract parameter, the party, and the class category;ranking, by the computing device, the contract based on the at least one contract parameter; and generating, on a user interface, a change recommendation for the contract including at least a correlation between a contractual element and the advisory recommendation.

10. The method of claim 9, further comprising: creating, by the computing device, a reduced size semantic equivalent contractual element for each of the contractual elements in the plurality of contractual elements.

11. The method of claim 9, further comprising: providing, by the computing device, a user selectable interface configured to enable selection of a subset of the contractual elements for analysis.12 The method of claim 9, further comprising: providing, by the computing device, a set of predefined prompts that are user selectable to assist a Large Language Model (LLM) to selectively review a subset of the contractual elements.

Images