What is this course about?
Legislative analysis is high-volume reading with a very low tolerance for error, and a summary that drops an operative clause can misinform a vote. This course covers bill tracking and change detection, summarization that preserves operative language, the limits of AI on fiscal numbers, and a verification protocol before anything reaches a decision-maker.
Who is this course for?
- Legislative analysts and policy staff
- Fiscal analysts and budget office staff
- Agency legislative affairs offices
- Committee staff
What do I need before starting?
- Familiarity with the legislative process in your jurisdiction
- No technical background required
What will I be able to do afterwards?
- Track bills and detect substantive change across versions
- Summarize legislation without losing operative language
- State clearly where AI must not touch fiscal numbers
- Run a verification protocol before anything reaches a decision-maker
- Maintain analytic neutrality in a politically sensitive tool
What does each module cover?
How do you track bills and detect real change?
45 minChange detection that distinguishes substantive amendment from renumbering.
Objectives
- Track bills across versions and sessions
- Distinguish substantive from technical change
- Alert on changes that matter to your portfolio
Topics
Activity. Build change detection for a real bill across three versions.
How do you summarize without losing operative language?
50 minSummarization that preserves the words that actually do the work.
Objectives
- Identify operative versus descriptive language
- Summarize while preserving operative terms
- Detect summaries that dropped something material
Topics
Activity. Summarize a complex bill and have a colleague find what the summary dropped.
Where must AI not touch the numbers?
50 minFiscal note support, and the hard line between narrative assistance and computed figures.
Objectives
- Separate narrative assistance from computation
- Require deterministic calculation for all figures
- Verify any number before it is published
Topics
Activity. Build a fiscal note workflow where every number comes from calculation, not generation.
How do you compare policy across jurisdictions?
45 minComparative analysis across states, with the sourcing that makes it defensible.
Objectives
- Compare comparable provisions across jurisdictions
- Source every comparison to primary text
- Handle jurisdictional differences honestly
Topics
Activity. Compare one policy across three states with citations to primary text.
How do you analyze constituent correspondence at volume?
45 minExtracting themes and positions from correspondence without misrepresenting constituents.
Objectives
- Extract themes and positions from high-volume correspondence
- Distinguish organized campaigns from individual contact
- Report without misrepresenting the input
Topics
Activity. Analyze a correspondence set and report themes with campaign detection.
What is the verification protocol?
40 minThe checks that must pass before analysis reaches a member or an executive.
Objectives
- Define a verification protocol with named steps
- Assign accountability for verification
- Make the protocol non-bypassable under time pressure
Topics
Activity. Write the verification protocol and test it under a simulated deadline.
How do you keep the tool analytically neutral?
40 minNeutrality and its appearance in a tool used by people with opposing positions.
Objectives
- Design for analytic neutrality
- Test for systematic framing bias
- Handle the perception of partisanship
Topics
Activity. Test your analysis tool for framing bias across politically opposed bills.
Building the bill analysis agent
50 minThe lab module: an agent that cites statutory text for every claim.
Objectives
- Build an agent with mandatory citation to primary text
- Verify every claim traces to the source
- Integrate the verification protocol
Topics
Activity. Build the agent and verify every claim in one analysis traces to statutory text.
What is the capstone project?
Bill analysis workflow with mandatory citation
Build a legislative analysis workflow with version change detection, operative-language-preserving summarization, a fiscal note process where numbers are computed not generated, a non-bypassable verification protocol, and an agent that cites statutory text for every claim.
Deliverable: A working analysis workflow with a completed, fully-cited analysis of a real bill.
How are learners assessed?
- Summary reviewed by a colleague for material omissions
- Every number in the fiscal work traced to a deterministic calculation
- Framing bias test run across politically opposed bills
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?
The hands-on modules run against agents already deployable on the ibl.ai platform for government.
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.
- Congress.gov
Library of Congress
Primary legislative text and version history used throughout.
- Government Accountability Office
GAO
Analytic standards and methodology reference.
- AI Guide for Government
GSA Centers of Excellence
Federal guidance on AI in analytic work.
- National Conference of State Legislatures
NCSL
State legislative process reference for the comparative module.
Delivery notes
Binding guidance for anyone preparing and delivering this course.
- Module 2's omission exercise is the course's most important. Participants consistently produce summaries that read well and drop an operative clause, and discovering it themselves is what builds the verification habit.
- Module 3's line must be absolute. Generated numbers in a fiscal note are a public credibility failure, and the workflow should make it structurally impossible rather than discouraged.
- Module 7's neutrality testing is delicate and necessary. Run it with bills the participants disagree about, or the test finds nothing.
- Legislative process differs by jurisdiction. Localize the tracking and fiscal modules rather than teaching the federal process as universal.
- Module 6 must survive time pressure. Verification protocols get bypassed during session, and the design should account for that rather than assume discipline.
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 Legislative and Policy Analysis with AI course cover?
Legislative analysis is high-volume reading with a very low tolerance for error, and a summary that drops an operative clause can misinform a vote. This course covers bill tracking and change detection, summarization that preserves operative language, the limits of AI on fiscal numbers, and a verification protocol before anything reaches a decision-maker. It runs 5 hours across 8 modules across 8 modules, at intermediate level, and closes with a capstone: Bill analysis workflow with mandatory citation.
Who should take Legislative and Policy Analysis with AI?
It is written for Legislative analysts and policy staff, Fiscal analysts and budget office staff, Agency legislative affairs offices, Committee staff. Prerequisites: Familiarity with the legislative process in your jurisdiction; 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 Legislative and Policy Analysis with AI?
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.