PRIMEMATCH AI™

AI TRANSPARENCY & RECOMMENDATION POLICY

Enterprise Partner Intelligence Platform

Version 1.0 — Draft for Attorney Review

Effective Date: February 2026

Last Updated: September 2026

Provider Legal Entity: Lyfe Share Inc., 6991 Peachtree Industrial Blvd. #400, Suite. A-15, Peachtree Corners, GA 30092

AI Governance Contact: michael@lyfeshare.io

IMPORTANT LEGAL NOTICE

This AI Transparency & Recommendation Policy is a comprehensive first draft prepared for business and legal planning and for review by qualified counsel. It is intended to operate together with the PrimeMatch AI™ Master Terms of Service, Privacy Policy, applicable Data Processing Addendum, Order Form, Service Level Agreement, marketplace terms, API terms, and other incorporated agreements. Bracketed provisions require completion or jurisdiction-specific review before publication or contractual use.

PrimeMatch AI™ provides artificial-intelligence-enabled informational and analytical decision-support. It does not replace professional judgment and does not make contracting, procurement, legal, financial, staffing, teaming, proposal, award, or marketplace decisions for users. Users remain solely responsible for reviewing outputs and making all final decisions.

TABLE OF CONTENTS

1. Purpose, Scope, and Policy Status

2. Definitions and Proprietary Intelligence Concepts

3. Core Transparency Principles

4. Human Responsibility and Decision-Support Boundary

5. AI System Architecture and Functional Categories

6. Data Inputs, Context, and Knowledge Sources

7. Recommendations, Rankings, Scores, and Confidence Values

8. Partner Intelligence and Teaming Recommendations

9. Opportunity Intelligence and Forecasting

10. Proposal Intelligence and AI-Generated Communications

11. Reputation Intelligence, Trust Score™, and Team Strength Index™

12. Explainability, Reason Codes, and User-Facing Disclosures

13. Limitations, Uncertainty, Errors, and Hallucination Risk

14. Human Review, Override, Appeals, and Feedback

15. Bias, Fairness, and Responsible Use

16. Government Contracting and Procurement-Specific Safeguards

17. Enterprise, White-Label, Multi-Tenant, and API Deployments

18. Marketplace Providers and Transactions

19. AI Agents™, Workflow Automation, and Autonomous Features

20. Training, Learning, Model Improvement, and Derived Intelligence

21. Security, Abuse Prevention, and Restricted Uses

22. Monitoring, Testing, Evaluation, and Change Management

23. Regulatory Transparency and Jurisdictional Adaptation

24. Intellectual Property, Confidentiality, and Proprietary Methodologies

25. Disclaimers; No Guarantees; Allocation of Responsibility

26. Limitation of Liability, Indemnification, and Remedies

27. Policy Administration, Updates, and Survival

28. Governing Law, Dispute Resolution, and General Provisions

Appendix A — AI Output Disclosure Matrix

Appendix B — Recommendation Review Checklist

Appendix C — Enterprise AI Governance Addendum Framework

Appendix D — User Acknowledgment of AI Decision-Support Responsibilities

1. PURPOSE, SCOPE, AND POLICY STATUS

1.1 Purpose.

This AI Transparency & Recommendation Policy (the “Policy”) explains how PrimeMatch AI™ uses artificial intelligence, machine learning, semantic technologies, knowledge graphs, analytical scoring, generative AI, workflow automation, and related computational methods to produce recommendations, rankings, scores, forecasts, summaries, analyses, communications, proposal assistance, and other outputs. The Policy is designed to provide meaningful transparency regarding the nature, intended use, limitations, and governance of such outputs while protecting PrimeMatch AI’s confidential and proprietary technology.

1.2 Scope.

This Policy applies to AI-enabled functions made available through PrimeMatch AI™, including multi-tenant SaaS environments, white-label enterprise portals, government-contractor environments, commercial and enterprise accounts, marketplace functions, APIs, integrations, executive dashboards, future mobile applications, and AI Agent™-enabled workflows. It applies whether an output is delivered directly to an end user, embedded in a customer workflow, exposed through an API, displayed in a white-label interface, or used to support another PrimeMatch AI™ feature.

1.3 Relationship to Other Agreements.

This Policy supplements and is incorporated by reference into the PrimeMatch AI™ Master Terms of Service where specified. If a conflict exists, the executed Order Form, negotiated enterprise agreement, Data Processing Addendum, or other agreement expressly designated as controlling shall govern to the extent of the conflict. This Policy does not independently create service-level commitments, warranties, or rights beyond those stated in the governing agreement.

1.4 No Professional Advice.

AI outputs are not legal, procurement, accounting, tax, investment, insurance, licensing, engineering, cybersecurity, personnel, or other regulated professional advice. PrimeMatch AI™ does not act as a contracting officer, procurement official, proposal evaluator, attorney, accountant, broker, insurer, guarantor, employment agency, or fiduciary.

2. DEFINITIONS AND PROPRIETARY INTELLIGENCE CONCEPTS

For purposes of this Policy, capitalized terms have the meanings below or in the Master Terms of Service. The following names identify proprietary software systems, methodologies, analytical frameworks, scoring systems, models, taxonomies, trademarks, trade secrets, know-how, and other intellectual property owned by or licensed to PrimeMatch AI. Their inclusion in this Policy does not disclose source code, model weights, proprietary formulas, ranking coefficients, feature engineering, trade secrets, or other confidential implementation details.

PrimeMatch AI™. the proprietary Partner Intelligence Platform and related services, applications, interfaces, APIs, systems, models, documentation, and technology.

PrimeMatch Confidence™. a proprietary analytical confidence framework that may communicate the degree of support, completeness, consistency, or relevance associated with an output; it is not a probability of award, success, honesty, solvency, compliance, or future performance.

Organization DNA™. a proprietary structured representation of an organization’s capabilities, qualifications, attributes, experience, preferences, constraints, differentiators, and related information.

Opportunity DNA™. a proprietary structured representation of an opportunity’s requirements, objectives, attributes, constraints, acquisition signals, evaluation factors, and related information.

Team DNA™. a proprietary representation of a proposed or existing team’s combined capabilities, coverage, complementarity, gaps, and relationship characteristics.

Relationship DNA™. a proprietary representation of relevant relationship history, interaction signals, collaboration context, and other permitted relationship information.

Proposal DNA™. a proprietary representation of proposal content, solicitation alignment, response structure, evidence, themes, gaps, and related analytical signals.

Growth DNA™. a proprietary framework for identifying business-development patterns, opportunities, readiness factors, and potential growth pathways.

Trust Score™. a proprietary decision-support score that may synthesize permitted reputation, verification, activity, relationship, performance, and other signals; it does not certify honesty, integrity, capability, compliance, or future performance.

Team Strength Index™. a proprietary decision-support measure of team composition and opportunity alignment; it is not a prediction or guarantee of proposal success or contract award.

PrimeMatch Intelligence™. the platform-wide proprietary intelligence layer encompassing multiple analytical and AI-enabled capabilities.

Partner Intelligence™. analytical functions supporting partner discovery, comparison, qualification, relationship assessment, and teaming analysis.

Opportunity Intelligence™. analytical functions supporting opportunity discovery, qualification, prioritization, forecasting, and alignment analysis.

Proposal Intelligence™. analytical and generative functions supporting proposal planning, analysis, drafting assistance, compliance review support, and content development.

Executive Intelligence™. business intelligence and decision-support functions designed for executive, portfolio, pipeline, operational, and strategic views.

Marketplace Intelligence™. analytical functions supporting marketplace discovery, provider comparison, transaction context, reputation signals, and marketplace decision-support.

Enterprise Intelligence™. cross-organizational intelligence, analytics, governance, and reporting functions available in enterprise environments.

AI Agent™. a software-based AI-enabled function that may perform or coordinate defined tasks, retrieve information, generate outputs, recommend actions, or initiate authorized workflow steps subject to configured permissions and safeguards.

Intelligence Engine™. a proprietary computational system that transforms inputs into structured intelligence, analyses, recommendations, or related outputs.

Semantic Matching Engine™. a proprietary matching system that evaluates semantic, contextual, attribute-based, and other signals to identify potential relationships among organizations, opportunities, capabilities, and teams.

Knowledge Graph™. a proprietary structured network of permitted entities, attributes, relationships, and contextual connections used to support search, analysis, inference, and decision-support.

Learning Engine™. a proprietary system that may use permitted feedback, usage patterns, evaluation results, and other data to improve models, rankings, workflows, and system performance subject to applicable agreements and privacy controls.

3. CORE TRANSPARENCY PRINCIPLES

3.1 Meaningful Notice.

PrimeMatch AI™ seeks to provide notice, in context and at a level appropriate to the use case, when a user is interacting with AI-enabled functionality or receiving material AI-assisted content, recommendations, rankings, scores, forecasts, or generated communications. Notice may appear through labels, tooltips, interface text, documentation, API metadata, policy links, or enterprise-configured disclosures.

3.2 Proportional Transparency.

Transparency does not require disclosure of confidential source code, model weights, security controls, proprietary algorithms, internal prompts, trade secrets, feature coefficients, model-routing logic, fraud controls, or information that would enable circumvention or misuse. PrimeMatch AI™ may provide category-level explanations, principal factors, reason codes, confidence indicators, source references when available, or other meaningful explanations without revealing protected implementation details.

3.3 Human Accountability.

Every recommendation, ranking, score, confidence value, prediction, forecast, summary, analysis, AI-generated communication, and proposal output is informational decision-support only. The user, customer, organization, marketplace participant, or other authorized human remains responsible for evaluating the output, obtaining appropriate professional advice, validating source information, and making any final business, procurement, legal, financial, staffing, contracting, or marketplace decision.

3.4 No Hidden Guarantee.

No confidence label, ranking position, score, forecast, recommendation, explanation, dashboard, badge, or other signal should be interpreted as a guarantee or certification of contract award, proposal success, business growth, profitability, successful partnership, legal compliance, licensing, insurance, financial stability, vendor performance, marketplace quality, partner honesty, compatibility, or any future outcome.

4. HUMAN RESPONSIBILITY AND DECISION-SUPPORT BOUNDARY

4.1 Decision-Support Only.

PrimeMatch AI™ is designed to augment human judgment by organizing information, surfacing relevant signals, identifying potential relationships, and generating analytical or drafting assistance. Unless expressly agreed in a separate written enterprise agreement and legally permissible, the platform is not intended to make legally binding, procurement, employment, credit, insurance, licensing, eligibility, award, debarment, suspension, source-selection, or other high-impact decisions on behalf of a user or customer.

4.2 User Review Obligations.

Before relying on an AI output, users should review the underlying context, confirm material facts, identify missing or stale information, consider contrary evidence, and evaluate whether the output is appropriate for the specific decision. Where an output concerns a solicitation, teaming arrangement, proposal, subcontract, marketplace provider, or regulatory matter, users are responsible for reviewing authoritative documents and obtaining appropriate professional advice.

4.3 No Delegation of Legal Duties.

Use of an AI Agent™, workflow automation, recommendation, or generated communication does not transfer to PrimeMatch AI™ any duty imposed on a user by law, regulation, contract, solicitation, procurement rule, corporate policy, professional standard, or fiduciary obligation. Users remain responsible for approvals, signatures, certifications, representations, submissions, records, and compliance determinations.

5. AI SYSTEM ARCHITECTURE AND FUNCTIONAL CATEGORIES

5.1 Functional Layers.

PrimeMatch AI™ may combine multiple computational approaches, including semantic retrieval, rules, statistical methods, machine learning, large language models, embeddings, vector search, knowledge graphs, ranking systems, classification, extraction, summarization, forecasting, and generative models. Different features may use different models or combinations of models and may change over time as described in Section 22.

5.2 Retrieval and Semantic Search.

Semantic search and retrieval features may identify content based on meaning, context, similarity, metadata, structured attributes, and relationships rather than exact keywords alone. Such systems may return relevant but non-identical results and may omit information that a user considers material.

5.3 Structured Intelligence.

The Intelligence Engine™, Semantic Matching Engine™, Knowledge Graph™, and related systems may convert raw or unstructured information into structured attributes, entity relationships, classifications, inferred signals, or analytical representations. Derived information may differ from a user’s own characterization and should be reviewed where material.

5.4 Generative AI.

Generative features may draft, summarize, transform, explain, compare, brainstorm, or otherwise produce text or other content based on user instructions, retrieved materials, model knowledge, organization context, opportunity context, and configured safeguards. Generated content may be incomplete, inaccurate, non-unique, outdated, or unsuitable for a particular submission or decision unless reviewed and revised by the user.

5.5 Forecasting and Predictive Analytics.

Forecasts and predictions may identify trends, potential timing, likelihood indicators, readiness signals, pipeline conditions, or other forward-looking analytical information. These outputs depend on assumptions and available data and may be affected by external events, procurement changes, policy changes, competitive activity, incomplete records, or factors not observable to PrimeMatch AI™.

6. DATA INPUTS, CONTEXT, AND KNOWLEDGE SOURCES

6.1 Input Categories.

AI-enabled features may process information supplied by users or customers, account and organization data, uploaded documents, capability statements, resumes, past-performance information, solicitation and opportunity data, marketplace information, interaction data, API-supplied data, administrative configurations, permitted public or licensed data, and other information described in the Privacy Policy or applicable agreement.

6.2 Source Quality.

PrimeMatch AI™ does not guarantee that third-party, public, customer-supplied, marketplace, or integrated information is complete, current, authentic, or error-free. A high-quality analytical method cannot eliminate defects in its inputs. Users should independently verify information that is material to a decision.

6.3 Derived and Inferred Information.

The platform may generate derived or inferred information, including structured attributes, semantic relationships, classifications, capability mappings, gap analyses, confidence indicators, recommendations, and other analytical outputs. Such information reflects computational assessment and should not be treated as an independently verified fact unless expressly identified as verified.

6.4 Source References.

Where technically and contractually feasible, certain features may display source references, provenance indicators, citations, data timestamps, or evidence links. The absence of a displayed source reference does not imply that an output is authoritative, and the presence of a source reference does not eliminate the user’s review obligations.

7. RECOMMENDATIONS, RANKINGS, SCORES, AND CONFIDENCE VALUES

7.1 Nature of Recommendations.

Recommendations identify potentially relevant actions, organizations, opportunities, providers, content, team compositions, or other options based on available signals and configured objectives. They are not endorsements, guarantees, certifications, or directives.

7.2 Rankings.

Ranking order may reflect multiple weighted and contextual factors and can vary by user, organization, opportunity, configuration, timing, data availability, or product version. A higher-ranked result is not necessarily superior in every respect and does not establish that a lower-ranked result is unsuitable.

7.3 Scores.

Scores, including PrimeMatch Confidence™, Trust Score™, Team Strength Index™, readiness scores, compatibility indicators, or similar measures, are proprietary decision-support signals. They may be normalized, categorical, comparative, contextual, or composite and should not be interpreted as objective truth, legally required ratings, or actuarial probabilities unless expressly stated otherwise in product documentation.

7.4 Confidence.

PrimeMatch Confidence™ may communicate the degree to which an output is supported by available data, model consistency, semantic alignment, completeness, or other proprietary factors. It is not a probability that a contract will be awarded, a proposal will succeed, a partner will perform, a provider is trustworthy, a business will grow, or any future outcome will occur.

7.5 Dynamic Outputs.

Recommendations and scores may change as data, models, system configurations, customer settings, feedback, opportunity information, or external conditions change. Historical values should not be assumed to remain valid indefinitely.

8. PARTNER INTELLIGENCE AND TEAMING RECOMMENDATIONS

8.1 Partner Discovery.

Partner Intelligence™ and the Semantic Matching Engine™ may identify organizations that appear relevant based on capabilities, certifications, experience, set-aside attributes, NAICS/PSC information, geography, relationship signals, opportunity requirements, availability, marketplace data, or other permitted attributes. A match does not establish qualification, responsibility, integrity, financial capacity, licensing, insurance, security clearance, organizational conflict status, eligibility, or willingness to team.

8.2 Organization DNA™ and Relationship DNA™.

Organization DNA™ and Relationship DNA™ may help structure and compare organizational and relationship information. Users must independently assess whether the information is accurate, current, complete, permitted for use, and sufficient for the contemplated teaming arrangement.

8.3 Team DNA™ and Team Strength Index™.

Team DNA™ and Team Strength Index™ may identify complementary capabilities, apparent gaps, overlap, risk factors, or alignment with opportunity requirements. These systems do not guarantee compatibility, successful negotiation, proposal performance, subcontract performance, compliance with limitations on subcontracting, organizational-conflict rules, small-business requirements, or any award outcome.

8.4 Due Diligence.

Users are responsible for partner due diligence, including verification of registrations, representations, certifications, exclusions, licenses, insurance, past performance, financial capacity, cybersecurity posture, conflicts, ownership, export-control implications, sanctions considerations, and other matters relevant to the transaction or procurement.

9. OPPORTUNITY INTELLIGENCE AND FORECASTING

9.1 Opportunity Analysis.

Opportunity Intelligence™ and Opportunity DNA™ may extract or infer requirements, deadlines, evaluation factors, customer needs, capability gaps, acquisition signals, likely teaming needs, strategic fit, and other attributes from available opportunity information. Users must review the official solicitation, amendment, notice, contract vehicle terms, and authoritative agency communications.

9.2 Forecasting.

Opportunity forecasting may estimate timing, fit, competition, readiness, pursuit priority, or related factors. Forecasts are inherently uncertain and may be affected by appropriations, agency priorities, procurement strategy, amendments, cancellations, protests, market conditions, and other events beyond PrimeMatch AI’s control.

9.3 No Procurement Authority.

PrimeMatch AI™ is not an official government procurement system, contracting office, source-selection authority, SAM.gov, agency forecast, or substitute for authoritative procurement sources. The platform does not determine eligibility, responsiveness, responsibility, evaluation results, or award status.

10. PROPOSAL INTELLIGENCE AND AI-GENERATED COMMUNICATIONS

10.1 Proposal Assistance.

Proposal Intelligence™ and Proposal DNA™ may assist with outlines, requirement mapping, compliance matrices, gap identification, content suggestions, summaries, drafts, themes, questions, red-team support, and other proposal-development tasks. Generated proposal content must be reviewed, edited, verified, and approved by qualified personnel before use or submission.

10.2 No Certification of Compliance.

A compliance-related indicator, checklist, or analysis is not a legal conclusion or certification that a proposal complies with a solicitation, regulation, statute, agency instruction, page limit, formatting rule, mandatory clause, representation, certification, or evaluation requirement. Users remain responsible for final compliance review.

10.3 AI-Generated Communications.

PrimeMatch AI™ may generate emails, introductions, partner outreach, questions, summaries, negotiation drafts, proposal language, executive communications, and other messages. Users are responsible for reviewing tone, accuracy, confidentiality, privilege implications, recipient suitability, legal effect, and authorization before sending or using such communications.

10.4 Human Authorship and Attribution.

Users should not misrepresent AI-generated content as independently verified or human-authored where disclosure is required by law, regulation, solicitation, customer policy, professional obligation, or applicable contract. PrimeMatch AI™ may provide or require AI-content notices where appropriate.

11. REPUTATION INTELLIGENCE, TRUST SCORE™, AND TEAM STRENGTH INDEX™

11.1 Reputation Intelligence.

Reputation Intelligence may aggregate or analyze permitted signals relating to public information, marketplace activity, platform interactions, verification status, relationship history, performance indicators, disputes, feedback, or other sources. Reputation Intelligence is contextual, may be incomplete, and does not constitute a background check or guarantee of character, honesty, integrity, performance, or future conduct.

11.2 Trust Score™.

Trust Score™ is a proprietary decision-support signal and not a credit score, consumer report, responsibility determination, suspension/debarment determination, security-clearance determination, or legal certification. Unless PrimeMatch AI™ expressly states otherwise in a separate written agreement, Trust Score™ is not intended for decisions governed by the Fair Credit Reporting Act or analogous consumer-reporting laws.

11.3 Team Strength Index™.

Team Strength Index™ evaluates selected team and opportunity alignment factors but does not account for every material issue. Users should consider contractual authority, negotiation dynamics, pricing, leadership, past performance, availability, competitive strategy, conflicts, customer relationships, and other factors that may not be represented in the score.

11.4 No Adverse Action Automation.

PrimeMatch AI™ should not be used to make fully automated adverse decisions regarding employment, credit, insurance, housing, education, essential services, or other regulated high-impact decisions unless expressly supported by the applicable product, enabled under a separate written agreement, and used in compliance with all applicable legal requirements.

12. EXPLAINABILITY, REASON CODES, AND USER-FACING DISCLOSURES

12.1 Explanation Methods.

Depending on the feature, PrimeMatch AI™ may provide explanations through principal-factor summaries, reason codes, matching attributes, source highlights, evidence references, confidence indicators, data-quality notices, missing-information flags, or natural-language explanations. Explanation formats may differ across products and model versions.

12.2 Limits of Explainability.

Complex AI systems may not permit a complete human-readable reconstruction of every computational step. PrimeMatch AI™ does not undertake to disclose proprietary source code, model weights, internal chain-of-thought, hidden reasoning traces, system prompts, security mechanisms, trade secrets, or other protected technical details. Where explanation is appropriate, PrimeMatch AI™ may provide a concise rationale, material factors, evidence, or functional description instead.

12.3 User-Facing Labels.

PrimeMatch AI™ may use labels such as “AI-assisted,” “AI-generated,” “recommendation,” “forecast,” “confidence,” “estimated,” “suggested,” “inferred,” or similar terminology to distinguish analytical or generated outputs from verified source facts. Enterprise customers may configure additional disclosures, provided they do not remove legally required notices or misrepresent the nature of the technology.

12.4 API Transparency.

API responses may include metadata, field names, model or feature identifiers, confidence values, explanation codes, source references, or documentation describing intended use and limitations. API customers are responsible for preserving legally required and contractually required disclosures when surfacing AI outputs to their own users.

13. LIMITATIONS, UNCERTAINTY, ERRORS, AND HALLUCINATION RISK

13.1 Inherent Limitations.

AI and machine-learning systems are probabilistic or inference-based and may produce incorrect, incomplete, inconsistent, biased, stale, non-responsive, or unexpected outputs. Generative models may produce plausible-sounding statements that are not supported by evidence (“hallucinations”). Semantic systems may misclassify or overlook context. Forecasts may fail. Recommendations may omit viable alternatives.

13.2 Material Verification.

Users must verify material factual statements, dates, citations, qualifications, certifications, clauses, regulatory requirements, pricing, technical specifications, past-performance assertions, security representations, and other consequential content before reliance or external use.

13.3 Confidence Is Not Accuracy Certification.

A high PrimeMatch Confidence™ value does not mean an output is error-free, complete, legally sufficient, or fit for a particular purpose. Confidence values should be considered together with source quality, context, recency, user expertise, and independent validation.

13.4 Third-Party Model Limitations.

Some features may rely on third-party AI, cloud, search, data, or infrastructure providers. PrimeMatch AI™ may implement safeguards, routing, evaluation, and contractual protections, but cannot eliminate all risks arising from third-party technologies or data sources.

14. HUMAN REVIEW, OVERRIDE, APPEALS, AND FEEDBACK

14.1 Review and Override.

Authorized users may disregard, modify, reject, override, or supplement recommendations unless a customer-configured workflow lawfully restricts that action. PrimeMatch AI™ does not require users to follow an AI recommendation merely because it is displayed or ranked highly.

14.2 Feedback Mechanisms.

The platform may allow users to rate, correct, flag, annotate, dispute, or provide feedback regarding outputs. Feedback may be used to improve the applicable user experience, customer configuration, evaluation process, or Learning Engine™ as permitted by the governing agreement and Privacy Policy.

14.3 Enterprise Review Channels.

Enterprise customers may establish internal review, escalation, approval, or exception processes for AI-assisted decisions. PrimeMatch AI™ may provide configurable permissions, audit logs, workflow controls, and administrative tools but does not assume responsibility for the customer’s internal governance decisions.

14.4 Correction Requests.

Where a user believes a material profile attribute, organization record, marketplace fact, or other underlying data is incorrect, the user should use available correction mechanisms or contact [SUPPORT / PRIVACY / AI GOVERNANCE CONTACT]. PrimeMatch AI™ may require verification before modifying information affecting other users, organizations, or marketplace participants.

15. BIAS, FAIRNESS, AND RESPONSIBLE USE

15.1 Risk Awareness.

PrimeMatch AI™ recognizes that AI systems can reflect limitations or bias in training data, public information, customer data, human feedback, labels, historical patterns, and model behavior. PrimeMatch AI™ may use testing, review, data-quality controls, evaluation, monitoring, and product safeguards to identify and reduce material risks appropriate to the use case.

15.2 No Unlawful Discrimination.

Users may not use PrimeMatch AI™ to unlawfully discriminate or to make prohibited decisions based on protected characteristics. Customers are responsible for configuring and using the platform consistently with applicable civil-rights, equal-opportunity, employment, procurement, and anti-discrimination requirements.

15.3 Proxy and Context Risk.

Business attributes such as location, industry, business size, contracting history, certifications, or relationship patterns may correlate with other characteristics. Users should evaluate whether reliance on any attribute is appropriate and lawful in the context of the decision.

15.4 Testing and Evaluation.

PrimeMatch AI™ may conduct quantitative and qualitative evaluations, red-team exercises, scenario testing, adversarial testing, human review, performance monitoring, or other assessments appropriate to the feature and risk profile. No testing process can demonstrate that an AI system will be error-free in all environments.

16. GOVERNMENT CONTRACTING AND PROCUREMENT-SPECIFIC SAFEGUARDS

16.1 Procurement Context.

PrimeMatch AI™ supports government contractors and other procurement participants but does not provide official procurement determinations. Users are responsible for compliance with applicable procurement statutes, regulations, agency supplements, solicitation terms, contract clauses, representations and certifications, ethics requirements, organizational-conflict rules, socioeconomic program rules, cybersecurity obligations, export controls, and other requirements.

16.2 FAR, DFARS, and Agency Rules.

PrimeMatch AI™ may reference or analyze Federal Acquisition Regulation (“FAR”), Defense Federal Acquisition Regulation Supplement (“DFARS”), agency supplements, clauses, policies, or solicitation provisions. Such assistance is informational only and does not guarantee that a user has identified every applicable requirement or interpreted it correctly. Users should consult authoritative sources and qualified counsel or contracting professionals.

16.3 Sensitive Procurement Information.

Users must not upload, request, disclose, or use source-selection information, contractor bid or proposal information, classified information, controlled unclassified information, export-controlled technical data, procurement-sensitive information, or other restricted data except where the specific PrimeMatch AI™ environment is expressly authorized in writing to receive such data and all required controls are in place.

16.4 Representations and Certifications.

PrimeMatch AI™ may help organize or draft information used in representations, certifications, responsibility determinations, capability statements, or proposals, but the user remains solely responsible for the accuracy, completeness, authorization, and legal effect of all submissions and certifications.

16.5 Procurement Integrity.

Users may not use AI-enabled functions to facilitate collusion, bid-rigging, improper information exchange, false claims, false statements, misrepresentation of status or past performance, evasion of competition requirements, procurement-integrity violations, or other unlawful conduct.

17. ENTERPRISE, WHITE-LABEL, MULTI-TENANT, AND API DEPLOYMENTS

17.1 Enterprise Configuration.

Enterprise customers may configure terminology, workflows, permissions, approval chains, data visibility, dashboards, recommendation criteria, integration settings, and white-label presentation. Customer configuration may materially affect the outputs users receive. The customer is responsible for ensuring its configurations are appropriate for its intended use and legal obligations.

17.2 White-Label Portals.

A white-label portal may present PrimeMatch AI™ functionality under customer branding. Unless otherwise disclosed, the underlying AI capabilities remain PrimeMatch AI™ technology. Enterprise customers may not remove mandatory AI notices or represent proprietary PrimeMatch AI™ systems as customer-owned technology except to the extent expressly authorized in writing.

17.3 Multi-Tenant Architecture.

PrimeMatch AI™ is designed as a multi-tenant service with logical access controls, organization hierarchies, role-based permissions, and other safeguards. Recommendations may be scoped by tenant, organization, user, role, dataset, or contractual configuration. A user should not assume that an output reflects information unavailable within the user’s authorized context.

17.4 APIs and Integrations.

Customers integrating PrimeMatch AI™ through APIs are responsible for appropriate authentication, authorization, downstream display, human-review controls, logging, rate limiting, user notices, and compliance in their own applications. PrimeMatch AI™ may require specific labels, attribution, metadata preservation, or disclosure language for certain AI outputs.

17.5 Customer-Specific Models and Data.

Where enterprise services include customer-specific retrieval, configuration, fine-tuning, evaluation, or other customization, ownership and permitted use of customer data and resulting configurations shall be governed by the applicable agreement. PrimeMatch AI™ retains all rights in its underlying platform, methodologies, models, improvements, generalized know-how, and proprietary intelligence systems except as expressly agreed otherwise.

18. MARKETPLACE PROVIDERS AND TRANSACTIONS

18.1 Marketplace Decision-Support.

Marketplace Intelligence™ may assist users in discovering, comparing, evaluating, or communicating with marketplace providers. Recommendations, reputation indicators, Trust Score™, rankings, or other signals do not constitute a warranty or endorsement of a provider or transaction.

18.2 Independent Providers.

Marketplace providers are independent parties unless expressly stated otherwise. PrimeMatch AI™ does not guarantee provider identity, licensing, insurance, financial condition, service quality, honesty, availability, performance, compatibility, or compliance. Users are responsible for diligence and contractual protections appropriate to the transaction.

18.3 Transaction Responsibility.

Unless expressly stated in a separate agreement, PrimeMatch AI™ is not a party to contracts between marketplace participants and is not responsible for performance, payment, delivery, disputes, warranties, or legal compliance arising from those transactions.

19. AI AGENTS™, WORKFLOW AUTOMATION, AND AUTONOMOUS FEATURES

19.1 AI Agent™ Functions.

AI Agents™ may be configured to gather information, summarize records, draft content, recommend actions, create task sequences, trigger authorized workflows, or interact with integrated systems. The scope of an AI Agent™ depends on its permissions, configuration, connected applications, user instructions, and product safeguards.

19.2 Authorization Boundaries.

AI Agents™ may not exceed the permissions granted to the applicable user, organization, or integration. Customers are responsible for assigning roles, approving connectors, managing credentials, and determining which actions require human approval.

19.3 Human-in-the-Loop Controls.

PrimeMatch AI™ may require confirmation before certain external communications, submissions, transactions, deletions, commitments, or other consequential actions. Customers should configure additional human approval where appropriate to their risk tolerance and legal obligations.

19.4 No Autonomous Contracting Authority.

Unless expressly agreed in writing and legally valid, no AI Agent™ has authority to bind PrimeMatch AI™, a customer, or a user to a contract, certify compliance, submit a proposal, make a procurement representation, accept an award, hire personnel, incur debt, transfer funds, or make another legally binding commitment.

20. TRAINING, LEARNING, MODEL IMPROVEMENT, AND DERIVED INTELLIGENCE

20.1 Learning Engine™.

The Learning Engine™ may use permitted feedback, evaluation data, corrections, usage signals, and other information to improve retrieval, ranking, matching, quality, safety, or user experience. Whether customer content may be used for generalized model training is governed by the Privacy Policy, Data Processing Addendum, enterprise agreement, and applicable product settings.

20.2 Customer Content.

PrimeMatch AI™ does not acquire ownership of customer content merely because it is processed by an AI feature. Rights in customer data, generated outputs, platform improvements, and derived or aggregated information are allocated under the governing agreement.

20.3 Proprietary Improvements.

Subject to contractual and privacy restrictions, PrimeMatch AI™ may develop generalized improvements, methods, evaluation techniques, taxonomies, templates, safety systems, ranking logic, and other know-how that do not disclose customer confidential information. Such improvements remain PrimeMatch AI intellectual property.

20.4 Model Change.

Underlying models, prompts, retrieval systems, ranking systems, safety controls, infrastructure providers, or other technical components may change over time. PrimeMatch AI™ may update systems to improve performance, security, reliability, cost, legal compliance, or functionality, subject to contractual commitments and change-management practices.

21. SECURITY, ABUSE PREVENTION, AND RESTRICTED USES

21.1 Security Controls.

PrimeMatch AI™ may use authentication, authorization, encryption, logging, tenant isolation, monitoring, rate limiting, content filtering, abuse detection, model safeguards, and other administrative, technical, and organizational measures appropriate to the service. No system is completely secure, and security commitments are governed by the applicable agreement and security documentation.

21.2 Prohibited Manipulation.

Users may not attempt to manipulate, poison, reverse engineer, extract, evade, or interfere with recommendation logic, safety controls, model behavior, Trust Score™, PrimeMatch Confidence™, Team Strength Index™, or other proprietary systems, except to the extent such restriction is prohibited by applicable law.

21.3 Prompt Injection and Malicious Content.

AI systems may be exposed to malicious instructions embedded in documents, websites, integrations, or other data. PrimeMatch AI™ may use safeguards to reduce prompt-injection and data-exfiltration risk, but users should exercise caution with untrusted content and should not rely on automated outputs as a substitute for security review.

21.4 Restricted Data.

Users must comply with applicable restrictions on personal data, health data, financial data, classified information, controlled unclassified information, export-controlled information, source-selection information, credentials, secrets, and other sensitive information. Product-specific documentation may impose additional restrictions.

22. MONITORING, TESTING, EVALUATION, AND CHANGE MANAGEMENT

22.1 AI Governance Program.

PrimeMatch AI™ may maintain governance practices appropriate to its services, including inventorying AI-enabled functions, documenting intended uses, evaluating risk, testing outputs, reviewing incidents, tracking material changes, monitoring performance, and assigning internal responsibilities. These practices may evolve with technology, law, and industry standards.

22.2 Testing and Evaluation.

PrimeMatch AI™ may use testing, evaluation, verification, validation, benchmarking, quality review, adversarial testing, red teaming, human evaluation, or other methods to assess system performance and risk. Evaluation results are context-dependent and do not guarantee future performance in every customer environment.

22.3 Material Changes.

PrimeMatch AI™ may update this Policy or provide additional notice when a material product change significantly alters the nature of AI interaction, the role of human review, the categories of consequential recommendations, or legally required transparency disclosures. Routine model updates, infrastructure changes, prompt refinements, or quality improvements may not require individualized notice unless required by law or contract.

22.4 Incident Review.

PrimeMatch AI™ may investigate reported AI incidents, systematic errors, misuse, safety concerns, or material output failures and may suspend, modify, restrict, or disable affected features as reasonably necessary to protect users, customers, third parties, or the platform.

23. REGULATORY TRANSPARENCY AND JURISDICTIONAL ADAPTATION

23.1 Applicable Law.

AI laws and regulations vary by jurisdiction and are evolving. PrimeMatch AI™ may implement disclosures, notices, labels, human-oversight controls, documentation, opt-outs, assessments, or other measures where required for a particular product, user population, or jurisdiction. This Policy should be reviewed by qualified counsel before publication in each material jurisdiction.

23.2 AI Interaction Disclosure.

Where applicable law requires users to be informed that they are interacting with an AI system, PrimeMatch AI™ may provide a clear notice through the applicable interface, workflow, or documentation. Enterprise and API customers are responsible for preserving such notices in downstream experiences when they are legally or contractually required.

23.3 Synthetic or AI-Generated Content.

Where applicable law requires machine-readable marking, labeling, or disclosure for AI-generated or manipulated content, PrimeMatch AI™ may provide or support appropriate mechanisms for covered features. Customers and users remain responsible for downstream uses that modify, remove, or republish such content.

23.4 High-Risk or High-Impact Uses.

PrimeMatch AI™ is not designed by default to make final high-impact decisions affecting legal rights or access to essential opportunities or services. A customer seeking to deploy the platform in a regulated high-risk or high-impact context must first obtain written authorization from PrimeMatch AI™ and is responsible for all assessments, notices, human review, appeals, documentation, and other legal obligations applicable to that deployment.

23.5 Jurisdiction-Specific Placeholder.

[COUNSEL REVIEW: Insert or cross-reference any jurisdiction-specific AI transparency, automated-decision, consumer-protection, procurement, employment, privacy, sectoral, or disclosure terms required for the jurisdictions in which PrimeMatch AI™ will be offered.]

24. INTELLECTUAL PROPERTY, CONFIDENTIALITY, AND PROPRIETARY METHODOLOGIES

24.1 Ownership.

As between PrimeMatch AI™ and the user or customer, PrimeMatch AI™ and its licensors retain all right, title, and interest in the platform and all related intellectual property, including PrimeMatch Confidence™, Organization DNA™, Opportunity DNA™, Team DNA™, Relationship DNA™, Proposal DNA™, Growth DNA™, Trust Score™, Team Strength Index™, PrimeMatch Intelligence™, Partner Intelligence™, Opportunity Intelligence™, Proposal Intelligence™, Executive Intelligence™, Marketplace Intelligence™, Enterprise Intelligence™, AI Agent™, Intelligence Engine™, Semantic Matching Engine™, Knowledge Graph™, Learning Engine™, algorithms, methods, taxonomies, ranking logic, scoring logic, prompts, system designs, workflows, models, model-routing techniques, documentation, interfaces, and improvements.

24.2 Transparency Does Not Waive Trade Secret Protection.

Nothing in this Policy requires PrimeMatch AI™ to disclose trade secrets, confidential algorithms, source code, model weights, hidden prompts, security measures, anti-fraud logic, proprietary training methods, internal evaluation datasets, or other information whose disclosure could impair security, intellectual property rights, or legitimate business interests.

24.3 Feedback.

To the extent permitted by the governing agreement, PrimeMatch AI™ may use feedback, suggestions, and non-confidential improvement ideas regarding AI features without restriction or compensation, provided that such use does not disclose customer confidential information or violate applicable data-protection obligations.

24.4 Third-Party Rights.

Certain AI models, data sources, software, or services may be provided by third parties and remain subject to their respective intellectual-property rights and contractual terms. PrimeMatch AI™ does not grant rights it does not possess.

25. DISCLAIMERS; NO GUARANTEES; ALLOCATION OF RESPONSIBILITY

25.1 Informational Outputs.

ALL AI-GENERATED OR AI-ASSISTED RECOMMENDATIONS, RANKINGS, SCORES, CONFIDENCE VALUES, PREDICTIONS, FORECASTS, SUMMARIES, ANALYSES, COMMUNICATIONS, PROPOSAL CONTENT, AND OTHER OUTPUTS ARE INFORMATIONAL DECISION-SUPPORT TOOLS ONLY. USERS REMAIN SOLELY RESPONSIBLE FOR ALL BUSINESS, PROCUREMENT, LEGAL, FINANCIAL, STAFFING, CONTRACTING, TEAMING, PROPOSAL, MARKETPLACE, AND OTHER DECISIONS.

25.2 No Outcome Guarantees.

PRIMEMATCH AI™ DOES NOT GUARANTEE CONTRACT AWARDS, PROPOSAL SUCCESS, BUSINESS GROWTH, PROFITABILITY, SUCCESSFUL PARTNERSHIPS, LEGAL COMPLIANCE, LICENSING, INSURANCE, FINANCIAL STABILITY, VENDOR PERFORMANCE, MARKETPLACE QUALITY, PARTNER HONESTY, COMPATIBILITY, OR FUTURE OUTCOMES.

25.3 AI and Data Disclaimer.

TO THE MAXIMUM EXTENT PERMITTED BY LAW AND EXCEPT AS EXPRESSLY PROVIDED IN A WRITTEN AGREEMENT, AI OUTPUTS ARE PROVIDED “AS IS” AND “AS AVAILABLE” AND MAY CONTAIN ERRORS, OMISSIONS, BIAS, INCONSISTENCIES, OR OUTDATED INFORMATION. PRIMEMATCH AI™ DISCLAIMS WARRANTIES THAT ANY AI OUTPUT WILL BE ACCURATE, COMPLETE, UNIQUE, CURRENT, UNINTERRUPTED, ERROR-FREE, OR SUITABLE FOR A PARTICULAR DECISION.

25.4 User Assumption of Decision Risk.

The user assumes responsibility for determining whether and how to use an output, for verifying material facts, for obtaining professional advice where appropriate, and for consequences arising from the user’s decision to rely on, modify, communicate, submit, or disregard an output.

26. LIMITATION OF LIABILITY, INDEMNIFICATION, AND REMEDIES

26.1 Incorporation of Master Terms.

The limitation-of-liability, exclusion-of-damages, indemnification, warranty-disclaimer, remedies, force-majeure, and risk-allocation provisions in the PrimeMatch AI™ Master Terms of Service or applicable negotiated agreement apply to this Policy and all AI-enabled features. This Section does not expand any liability cap or remedy stated in the governing agreement.

26.2 AI-Specific Allocation.

Without limiting the governing agreement, PrimeMatch AI™ shall not be liable for decisions made by users or customers based on AI outputs; failure to verify output; use of generated content in proposals, certifications, representations, contracts, or communications; customer configuration; downstream API implementation; marketplace transactions; or third-party data or model errors, except to the extent liability cannot lawfully be excluded.

26.3 User Indemnification.

To the extent provided in the governing agreement, the user or customer shall defend, indemnify, and hold harmless PrimeMatch AI™ and its affiliates from third-party claims arising out of unlawful or unauthorized use of AI features, misuse of outputs, prohibited discrimination, false submissions, procurement-integrity violations, infringement caused by user-provided content or instructions, or failure to comply with legally required notices or human-review obligations in a downstream deployment.

26.4 Jurisdictional Review.

[COUNSEL REVIEW: Confirm enforceability, caps, exclusions, indemnity scope, public-sector exceptions, consumer-law carve-outs, and any insurance requirements under the selected governing law and customer segments.]

27. POLICY ADMINISTRATION, UPDATES, AND SURVIVAL

27.1 Administration.

PrimeMatch AI™ may designate an AI governance, legal, compliance, security, product, or other function to administer this Policy. Questions may be directed to [AI GOVERNANCE CONTACT].

27.2 Updates.

PrimeMatch AI™ may amend this Policy to reflect changes in technology, product functionality, law, regulation, industry practice, risk management, or business operations. Material changes will be communicated as required by applicable law or contract. The effective date above identifies the current published version.

27.3 Versioning.

PrimeMatch AI™ may maintain archived versions or effective-date records for enterprise governance, audits, legal review, or customer administration. Enterprise customers may receive additional change notices under negotiated agreements.

27.4 Survival.

Provisions concerning intellectual property, confidentiality, disclaimers, allocation of responsibility, limitations of liability, indemnification, dispute resolution, and any other provisions that by their nature should survive shall survive termination or expiration to the extent stated in the governing agreement.

28. GOVERNING LAW, DISPUTE RESOLUTION, AND GENERAL PROVISIONS

28.1 Governing Law.

This Policy shall be governed together with the Master Terms of Service by the law specified in the applicable governing agreement. [DEFAULT GOVERNING LAW: STATE OF ______, WITHOUT REGARD TO CONFLICT-OF-LAWS PRINCIPLES.]

28.2 Dispute Resolution.

Any dispute arising from this Policy shall be resolved under the dispute-resolution provisions of the Master Terms of Service or applicable negotiated agreement, including any arbitration, venue, jury-trial waiver, class-action waiver, or public-sector exception stated there. [COUNSEL REVIEW REQUIRED.]

28.3 Severability.

If any provision of this Policy is held unenforceable, it shall be enforced to the maximum extent permitted and the remaining provisions shall remain in effect, subject to the governing agreement.

28.4 No Third-Party Beneficiaries.

Except as expressly provided in the governing agreement, this Policy creates no third-party beneficiary rights.

28.5 Entire AI Transparency Statement.

Together with incorporated agreements, product-specific disclosures, and legally required notices, this Policy states the general transparency and recommendation-governance framework for PrimeMatch AI™. Product-specific disclosures may provide more detailed information for particular features.

APPENDIX A — AI OUTPUT DISCLOSURE MATRIX

The following matrix illustrates the minimum intended transparency posture for principal PrimeMatch AI™ output categories. Product-specific implementation may vary, and legally required disclosures control.

APPENDIX B — RECOMMENDATION REVIEW CHECKLIST

Before making a consequential decision based in whole or in part on a PrimeMatch AI™ output, users should consider the following review steps as applicable. This checklist is informational and does not replace a customer’s internal controls or professional advice.

• Confirm that the underlying organization, opportunity, partner, proposal, or marketplace data is current and complete.
• Identify whether the output is a verified fact, user-supplied fact, derived inference, recommendation, score, forecast, or generated content.
• Review displayed sources, evidence, reason codes, confidence indicators, timestamps, and missing-data notices.
• Compare the output against authoritative solicitation, contract, agency, legal, regulatory, or customer sources where applicable.
• Consider alternative partners, opportunities, team structures, interpretations, and scenarios not surfaced by the recommendation.
• Assess whether any material factor may be absent, stale, biased, disputed, or outside the system’s authorized data context.
• Perform appropriate partner, vendor, marketplace, financial, cybersecurity, licensing, insurance, eligibility, and reputation due diligence.
• Obtain legal, procurement, financial, accounting, security, or other professional review where the decision warrants it.
• Confirm that any generated proposal, representation, certification, or communication is accurate, authorized, and appropriate before submission or transmission.
• Document the responsible human decision-maker and any required internal approval, override, or exception process.
• Do not interpret PrimeMatch Confidence™, Trust Score™, Team Strength Index™, rankings, or forecasts as guarantees or certifications.
• Report material suspected errors or systemic issues through the available feedback or support channel.

APPENDIX C — ENTERPRISE AI GOVERNANCE ADDENDUM FRAMEWORK

An enterprise customer may request or negotiate a separate AI Governance Addendum addressing the following matters where commercially and legally appropriate. Unless executed by authorized representatives, this Appendix does not create additional contractual obligations.

• Authorized AI use cases and expressly prohibited use cases;
• Customer-specific human-review and approval requirements;
• Permitted data categories and restricted-data handling;
• Customer-specific model, retrieval, or configuration controls;
• AI inventory, documentation, and change-notification commitments;
• Testing, evaluation, validation, or audit-right parameters;
• Incident notification and escalation processes;
• Bias, fairness, and high-impact-use controls;
• API and downstream transparency obligations;
• White-label attribution and disclosure requirements;
• Customer-specific retention, logging, and evidence requirements;
• Public-sector procurement, records, security, and sovereign-immunity considerations;
• International AI-regulation responsibilities;
• Allocation of responsibility for customer-built workflows and downstream decisions;
• Any negotiated service levels, warranties, liability allocations, or insurance requirements.

APPENDIX D — USER ACKNOWLEDGMENT OF AI DECISION-SUPPORT RESPONSIBILITIES

By using AI-enabled functions of PrimeMatch AI™, and subject to the governing agreement, the user acknowledges and agrees that:

• PrimeMatch AI™ uses artificial intelligence and related computational technologies to assist with analysis, matching, recommendations, rankings, scoring, forecasting, content generation, and workflow support.
• AI outputs may contain errors, omissions, uncertainty, bias, or outdated information and must be reviewed before consequential use.
• PrimeMatch AI™ provides decision-support and does not make final business, procurement, legal, financial, staffing, contracting, teaming, proposal, or marketplace decisions for the user.
• PrimeMatch Confidence™, Trust Score™, Team Strength Index™, and other scores or rankings are proprietary analytical signals and are not guarantees, certifications, or probabilities of future success.
• The user remains responsible for due diligence, professional advice, authoritative-source review, regulatory compliance, approvals, representations, certifications, submissions, and legally binding actions.
• PrimeMatch AI™ does not guarantee contract awards, proposal success, business growth, profitability, successful partnerships, legal compliance, licensing, insurance, financial stability, vendor performance, marketplace quality, partner honesty, compatibility, or future outcomes.
• The user will not use AI-enabled features for unlawful discrimination, procurement misconduct, false statements, unauthorized disclosure of restricted information, or other prohibited conduct.
• The user will preserve required AI disclosures and human-review safeguards when integrating PrimeMatch AI™ outputs into downstream systems or workflows.

Enterprise Customer / User Acknowledgment (if separately required):

Name: ________________________________ Title: ________________________________

Organization: ___________________________ Date: ________________________________

Signature: ______________________________

END OF PRIME MATCH AI™ AI TRANSPARENCY & RECOMMENDATION POLICY v1.0

Draft for Attorney Review — February 2026