# AI in Public Procurement: Writing an RFP That Gets Real Bids > Government · AI Course · GOV-4 > Source: https://ibl.ai/solutions/government/course/ai-in-public-procurement > Last updated: 2026-08-25 **Specify AI in a solicitation so you get comparable, honest proposals — required disclosures, evaluation criteria, and contract terms that preserve agency control.** ## The Short Answer **Generic AI solicitations produce proposals that cannot be compared, because vendors answer different questions. ibl.ai meets the terms agencies should require — source code and weights delivered, no per-seat pricing, and you own all the code and the data, so exit provisions are real rather than a migration services quote.** On ibl.ai you own all the code and the data, run it model-agnostic across any LLM, and pay with no per-seat pricing — so you can deploy anywhere, from your own cloud to a fully air-gapped network. [Request Access](https://ibl.ai/contact) · [Explore Government](https://ibl.ai/solutions/government) ## Course facts - **Level:** Intermediate - **Duration:** 5.5 hours across 8 modules - **Format:** Cohort workshop producing solicitation language - **Modules:** 8 - **Catalog code:** GOV-4 - **Frameworks covered:** FAR, FedRAMP, NIST AI RMF, State procurement law ## What is this course about? Generic AI solicitations attract proposals that cannot be compared. This course builds requirements that separate capability from marketing, mandatory disclosures on models and data handling, scoreable evaluation criteria, and the data rights and exit provisions that determine whether the agency stays in control after award. ## Who is this course for? - Contracting officers and specialists - Program managers writing requirements - Technical evaluation panel members - Agency counsel supporting acquisition ### What do I need before starting? - Familiarity with your agency's acquisition process - No technical background required ## What will I be able to do afterwards? - Write requirements that separate capability from vendor marketing - Mandate the disclosures needed to evaluate an AI system - Build evaluation criteria that are genuinely scoreable - Compare per-seat against usage and owned pricing in a solicitation - Secure data rights, code escrow, and exit provisions ## What does each module cover? ### Module 1 — Why do generic AI solicitations fail? How vague requirements produce incomparable proposals and a protest-prone evaluation. _(40 min)_ **Objectives** - Diagnose why a solicitation produced unusable proposals - Identify requirement language that invites marketing - Set specificity standards **Topics:** Requirement vagueness · Incomparable proposals · Protest risk · Specificity standards **Activity:** Critique a real AI solicitation for language that produced marketing responses. ### Module 2 — How do you specify capability rather than features? Performance-based requirements that a vendor must demonstrate rather than assert. _(50 min)_ **Objectives** - Write performance-based capability requirements - Specify demonstrations rather than claims - Avoid requirements that lock in one architecture **Topics:** Performance-based requirements · Demonstration requirements · Architecture neutrality · Measurability **Activity:** Rewrite five feature requirements as demonstrable capability requirements. ### Module 3 — What must a vendor disclose? Models, training data, subprocessors, and residency — the disclosures without which evaluation is guesswork. _(50 min)_ **Objectives** - Mandate the disclosures needed for evaluation - Require them in a comparable format - Handle claimed proprietary exemptions **Topics:** Model disclosure · Training data provenance · Subprocessors · Data residency **Activity:** Draft the mandatory disclosure section with a required response format. ### Module 4 — How do you build criteria that are actually scoreable? Evaluation criteria a panel can apply consistently and defend in a protest. _(50 min)_ **Objectives** - Write criteria that produce consistent scores - Weight criteria before receiving proposals - Document the evaluation defensibly **Topics:** Scoreable criteria · Weighting · Panel consistency · Protest defensibility **Activity:** Build the evaluation rubric and test panel consistency on a sample proposal. ### Module 5 — How do you compare pricing shapes fairly? Structuring the price schedule so per-seat, usage, and owned proposals can be compared at all. _(45 min)_ **Objectives** - Structure a price schedule that normalizes across shapes - Require lifecycle rather than year-one pricing - Model cost at realistic usage **Topics:** Price schedule design · Shape normalization · Lifecycle pricing · Usage assumptions **Activity:** Design a price schedule that makes three pricing shapes comparable. ### Module 6 — What data rights must the agency keep? Data rights, code escrow, and the terms that determine post-award control. _(45 min)_ **Objectives** - Specify required data and software rights - Determine when escrow is warranted - Protect agency data from vendor model training **Topics:** Data rights · Software rights · Escrow · Training prohibitions **Activity:** Draft the data rights and training prohibition clauses. ### Module 7 — How do you avoid a captive renewal? Pilot-to-production terms that keep a real competitive option at the next decision point. _(45 min)_ **Objectives** - Structure pilot terms that preserve competition - Require exit assistance and data return - Estimate switching cost before award **Topics:** Pilot structuring · Exit assistance · Data return · Switching cost **Activity:** Draft the exit provisions and estimate switching cost for a proposed award. ### Module 8 — Building the solicitation section The workshop module: a complete AI solicitation section with an evaluation rubric. _(55 min)_ **Objectives** - Assemble the complete solicitation section - Attach the evaluation rubric - Review with contracting and counsel **Topics:** Section assembly · Rubric attachment · Counsel review · Publication readiness **Activity:** Assemble the section and have a contracting officer review it. ## What is the capstone project? **AI solicitation section with an evaluation rubric.** Produce a complete solicitation section: performance-based capability requirements, mandatory vendor disclosures with a response format, a scoreable weighted evaluation rubric, a normalized price schedule, data rights and training prohibitions, and exit provisions. _Deliverable:_ A solicitation section reviewed by a contracting officer and agency counsel. ## How are learners assessed? - Panel consistency test — two evaluators must score a sample proposal similarly - Requirements checked for architecture neutrality - Exit provisions reviewed for whether they produce a real option ## What ships with the course? - **Facilitator guide.** Session-by-session running order, discussion prompts, and the questions that reliably derail a room. - **Learner workbook.** Exercises, checklists, and the templates each module's activity produces. - **Hands-on lab environment.** A sandboxed ibl.ai deployment so exercises run against real agents, not screenshots. - **Assessment bank.** Scenario questions and rubric criteria mapped to each stated learning outcome. - **Source bibliography.** Every primary regulation and standard cited on this page, linked and dated. ## Which AI agents does this course use? - [Procurement Agent](https://ibl.ai/solutions/government/agent/procurement-agent) - [Compliance Agent](https://ibl.ai/solutions/government/agent/compliance-agent) - [Budget Agent](https://ibl.ai/solutions/government/agent/budget-agent) - [Knowledge Agent](https://ibl.ai/solutions/government/agent/knowledge-agent) ## Where does the course material come from? Every module is grounded in primary sources — the regulation, standard, or research itself, not a summary of it. Each was resolved at authoring time. - [Federal Acquisition Regulation](https://www.acquisition.gov/far/) — Acquisition.gov. The regulatory framework governing the solicitation language. - [AI Guide for Government](https://coe.gsa.gov/coe/ai-guide-for-government/) — GSA Centers of Excellence. Federal acquisition practice guidance for AI. - [FedRAMP](https://www.fedramp.gov/) — FedRAMP PMO. Authorization requirements that must be reflected in the solicitation. - [AI Risk Management Framework](https://www.nist.gov/itl/ai-risk-management-framework) — NIST. Risk requirements referenced in the mandatory disclosures. ## Delivery notes Binding guidance for anyone preparing and delivering this course: - Module 4's panel consistency test is the protest-prevention exercise. Criteria that two evaluators score differently are criteria that will be protested. - Every clause must be reviewed by a contracting officer before delivery. Solicitation language written by technologists creates acquisition problems even when the technical content is right. - Module 2 must stay architecture-neutral. Requirements that only one vendor's design can satisfy invite protest regardless of whether that design is best. - State and local procurement differ substantially from federal. Localize the module for non-federal cohorts rather than assuming the FAR. - Do not write clauses that only ibl.ai can meet. The course's credibility with contracting officers depends on the requirements being genuinely neutral, and neutral requirements already favor an owned deployment on the merits. ## Why run AI training on a platform you own? - **You own the course, not a licence to it.** Course content, learner data, and the platform run inside your perimeter — you own all the code and the data. - **Model-agnostic delivery.** Run the course's AI components on any LLM — Claude, GPT, Llama, Gemini, Command — and switch anytime. - **No per-seat training licences.** Usage-based or self-hosted, so cost tracks actual use rather than headcount. - **Deploy anywhere.** Cloud, private VPC, on-premise, or fully air-gapped — including for cohorts that cannot use public AI tools. ## Frequently asked questions ### What does the AI in Public Procurement: Writing an RFP That Gets Real Bids course cover? Generic AI solicitations attract proposals that cannot be compared. This course builds requirements that separate capability from marketing, mandatory disclosures on models and data handling, scoreable evaluation criteria, and the data rights and exit provisions that determine whether the agency stays in control after award. It runs 5.5 hours across 8 modules across 8 modules, at intermediate level, and closes with a capstone: AI solicitation section with an evaluation rubric. ### Who should take AI in Public Procurement: Writing an RFP That Gets Real Bids? It is written for Contracting officers and specialists, Program managers writing requirements, Technical evaluation panel members, Agency counsel supporting acquisition. Prerequisites: Familiarity with your agency's acquisition process; No technical background required. ### Can we run this course on our own infrastructure? Yes. ibl.ai is model-agnostic and deploy-anywhere — cloud, private VPC, on-premise, or fully air-gapped — and you own all the code and the data. Cohort data, submissions, and any material learners upload stay inside your perimeter, which matters for government teams that cannot send work to a public AI tool. ### How do we get access to AI in Public Procurement: Writing an RFP That Gets Real Bids? Request access and we will set it up for your cohort — hosted by ibl.ai, or running against your own deployment. Tell us the group size and timing you need, and whether it should run inside your own perimeter. ### How much does AI training for government cost on ibl.ai? There is no per-seat pricing — you pay for usage or self-host and pay only for the infrastructure, so a 5,000-person rollout does not cost 5,000 licences. 1.6M+ users across 400+ organizations run the platform this way, including NVIDIA, MIT, and Syracuse University. ## More Government courses - [Getting an AI System Authorized: FedRAMP and NIST 800-53](https://ibl.ai/solutions/government/course/authorizing-ai-fedramp-nist-800-53): The authorization path for AI in a federal or state agency — control selection, boundary definition, and why an LLM complicates the system security plan. - [Sovereign and Air-Gapped AI for Public Agencies](https://ibl.ai/solutions/government/course/sovereign-air-gapped-ai-public-agencies): Run capable AI with no internet egress — model selection, air-gapped update paths, and the operational realities of a disconnected deployment. - [Citizen Service Agents: Design, Escalation, and Accessibility](https://ibl.ai/solutions/government/course/citizen-service-agents): Public-facing AI where the user has no alternative provider — plain language, Section 508 conformance, language access, and escalation that never traps a constituent. - [Records, FOIA, and AI: Retention When an Agent Writes](https://ibl.ai/solutions/government/course/records-foia-and-ai): What happens to public records law when an AI drafts the memo — retention, prompt logs as records, and responding to a request that reaches an AI system. - [AI Governance Inside a Public Agency](https://ibl.ai/solutions/government/course/ai-governance-public-agency): Stand up an agency AI governance program — inventory, use-case review board, impact assessment, and public transparency reporting. - [Legislative and Policy Analysis with AI](https://ibl.ai/solutions/government/course/legislative-policy-analysis-with-ai): Bill tracking, fiscal note support, and comparative policy research — with the verification discipline that keeps a wrong summary out of a member's briefing.