Legislative View.

A research service for legislative intelligence — public records, sourced

Where money meets policy, before it’s news.

Legislative View tracks the chain from IRS Form 990 grants to state-committee bill positions to cross-state template spread — surfacing the coordinated campaigns behind state legislation weeks before the press release.

The current corpus indexes 7,494 bills across all 50 states — full text and amendment history across 16,919 versions — joined to $98M in tracked grants to 313 recipient nonprofits from IRS Form 990 Schedule I filings, and 4,409 official position records filed by 2,147 distinct organizations in state committee proceedings.

i. The methodology

Three things your bill-tracking CRM can’t generate from bill text.

Public records have always held the answer. The work is the joining — federal 990 filings to state bill rosters to state-source position evidence. We do those joins continuously, against a database we control, with every join provable from a primary document.

First.

Follow the money.

We ingest every line item from IRS Form 990 Schedule I — grantor, recipient, amount, year — and join it through our organization registry to position records on bills you watch. Named grants are traced end-to-end, from filing image to bill testimony, with no inference between the steps.

Second.

See the coalition.

4,409 official position records from state committee proceedings. Coalition maps aggregate organizations across every in-scope bill — ranked by sustained engagement, not headline volume — so single-shot testifiers don’t crowd out the operators with durable interest.

Third.

Track model legislation.

When boilerplate from one state shows up in three more, we surface it the same week. Pairwise sentence-transformer scoring across all 50 states — with an explicit cross-state filter so within-state sibling-bill paraphrases don’t drown out the actual template spread.

The valuable work isn’t summarizing bill text — everyone can do that now. The valuable work is connecting bills to the organizations testifying on them, those organizations to the funders behind them, and those bills to similar bills in other states. The thesis of the platform

ii. A specimen

See the deliverable, not a mockup.

Open a real Legislative View brief — no signup wall.

“53 in-scope bills are active in California, with 1,917 official position records on file and a coalition map dominated by TechNet (11 bills opposed) on the industry side and Oakland Privacy (11 bills supported) on the civil-society side. The most-active bill, AB 1018 on automated decision systems, has 441 documented positions — 185 in support and 256 in opposition. Texas Appleseed (EIN 74-2804268) granted $10,000 to Consumer Federation of America in 2023; CFA subsequently testified in support of AB 1018.” Excerpted from the California brief — every claim has a numbered footnote linking to its primary source.

PDF · ~80 KB · 15 pages · California · ai-regulation · refreshed nightly

iii. Audit any claim

Pick any claim in that brief. We'll show you the source.

Quorum's AI Bill Tracker shows a "97% relevance" badge with no way to inspect what the model considered, and ships LLM-generated AI bill summaries no one can audit. We do the opposite — every number, every position, every grant traces to a primary document. Three examples pulled from the brief above:

  1. “CA AB 1018 has 441 documented positions on file — 185 support, 256 oppose.”

    Source: CA Assembly & Senate committee analyses leginfo.legislature.ca.gov/…?bill_id=202520260AB1018
  2. “Texas Appleseed (EIN 74-2804268) granted $10,000 to Consumer Federation of America in 2023.”

    Source: IRS Form 990 Schedule I, tax year 2023 projects.propublica.org/nonprofits/organizations/742804268
  3. “WA HB 2225 and WI AB 965 share 81% text similarity on AI chatbot regulation.”

    Source: Pairwise sentence-transformer cosine all-MiniLM-L6-v2, 384-dim cosine similarity

Every cell in the brief works this way. Pick any number — we'll send the SQL.

Every buyer who has been burned by a hallucinated AI summary starts asking the same question: where did this number come from? That question is our entire product. For the analyst, the journalist, the program officer

iv. Built for whom

Built for the buyers a legislative CRM can’t serve well.

Quorum, Phone2Action, and FiscalNote serve in-house GR teams running outreach. We serve the researchers and analysts a step removed — buyers whose job is to understand the policy landscape before anyone in it gets contacted.

Investigative journalists

Cite the funder network in a publishable piece.

A WSJ-grade investigation needs a paper trail from grant to recipient to bill testimony. A CRM’s email outbox doesn’t help. A footnoted brief with the IRS 990 image URL on every grant does.

“Show me where this number came from.”

Hedge-fund policy analysts

Predict bill spread before the market does.

When model legislation from a single state appears in three more, the affected sector reprices a week later. Template-similarity detection across all 50 states is the alpha. No CRM ships this.

“The risk isn’t the bill we know about — it’s the one we don’t.”

Foundation program staff

Understand the coalition before granting into it.

Before a foundation funds an advocacy organization, the program officer wants the coalition map — who else is in this fight, on which side, with what bill record. A CRM’s contact directory is overkill; the intelligence layer is the answer.

“Where would this grant actually land?”

State AGs & policy staff

Spot the model-legislation trend in your state.

When the same boilerplate that lost in California shows up in your state two months later, the lead time to draft an amicus or counterproposal matters. Cross-state similarity scoring is built for this. A CRM is the wrong product entirely.

“Is this our problem, or someone else’s problem we’re about to inherit?”

v. Inquiry

Tell us the question you’re trying to answer.

We’ll come back within one business day with a 20-minute walkthrough on your states, your bills, and the funder graph behind them.