Texas Legislature · Behavioral Intelligence

Know exactly where every Texas legislator stands.

Perch is an AI built on everything Texas legislators say on the record — committee testimony, official statements, press releases. Run any member, bill, or issue and get where they actually stand, cited to the statement that proves it.

Walk in knowing what they'll say.

Request Access →
See it live · no signup

You're one search away from knowing what they'll say.

A member, a committee, a strategy — the same engine reads all of it and answers from the record. Try one:

Or try one of these
A member
Know how anyone operates.
A behavioral profile — how she operates, votes, and who she sides with.
Give me a profile on Lois Kolkhorst
A committee
Read the room before you walk in.
How the room runs — throughput, who drives it, how it votes.
How does Senate Finance operate?
A strategy
Find who’s actually in play.
The play — who to target and why, straight from the record.
Who’s most persuadable on school choice?
The answer · example

A question in. A cited answer back.

Thousands of attributed statements resolve into one answer — every line traceable to the statement it came from.

Committee testimony
Floor vote · SB 827
Press remarks
Statement · Apr 23
Committee Q&A
→ The read
Where does Sen. Kolkhorst stand on rural water supply?
A consistent advocate for rural and small-system water infrastructure — focused on funding access for underserved counties.
Drawn from 12 attributed statements and 4 votes across committee and floor.
Sen. KolkhorstWater & Rural AffairsMay 6, 2026supportive
“…the smaller systems are the ones getting left behind, and the funding formula has to account for that…”
Sen. KolkhorstFinanceApr 23, 2026supportive
“…a county of nine thousand people has to be able to access these dollars…”
◆ every line pinned to the statement it came from
Every statement, read.
Every member, profiled.
Every claim, cited.
89th Regular + Interim · complete through Jul 30, 2026

Primary Sources Only

What they said. How they voted. Where they stand.

Member intelligence
Profiles built the way intelligence analysts read world leaders — from their own words. Perch studies everything a member has said in committee across all 182 legislators, and maps how they question, where they push back, and who they line up with.
What any member said
Find exactly what any member said about your issue — verbatim, dated, attributed to the hearing.
Committee vote record
See how members voted on every committee roll call — and where their votes diverged from their statements.
The profiles · example

Explain anyone in the Texas Legislature.

Sen. Charles Perry
R · SD-28 · 9 BILLS · n=214
Votes like they talk88%
Questions witnesses vs. speechmakinghigh
Party-line alignment71%
Top issue — water & rural affairslead
EVERY NUMBER CARRIES ITS n. BELOW MINIMUM SAMPLE STAYS BLANK, NEVER GUESSED.
Click any member and get the read: a behavioral fingerprint built from their own words and votes — how they question witnesses, whether they vote like they talk, which organizations they side with, and what they choose to file.
Methodology

Every measure follows a published method.

Each behavioral read is a named technique from political-science and NLP research — not a vibe — applied to the committee record and reported with its sample size.

01
Transcribe
committee transcripts + written record
02
Embed
1,536-d vectors, per statement
03
Classify
question & stance typing
04
Score
vs. all 181 peers + votes
05
Profile
6-dimension fingerprint, cited
Vector-space scaling
Voice fingerprint
Every statement becomes a 1,536-d embedding; each member’s centroid is scored by cosine similarity against all 181 peers to find who they sound most like.
Basis: word-embedding text scaling — Rheault & Cochrane, Political Analysis (2020) · n = 189,316
Leadership Trait Analysis
Behavioral profile
The framework used to profile heads of state from their public speech, applied to each member’s full committee record — six dimensions, each scored.
Basis: Leadership Trait Analysis — Hermann · 182 members
Question-function typology
Questioning style
Every question a member asks is classified — inquiry, position-taking, or challenge — and by who it’s aimed at.
Basis: rhetorical role of questions (Zhang et al., EMNLP 2017); message politics in committee hearings (Park, J. of Politics 2021) · n = 35,428
Speech-vs-vote divergence
Say-vs-vote consistency
Stance-classified statements aligned against recorded votes, per bill and venue — where the words and the vote agree, and where they don’t.
Basis: speech-vs-vote alignment · gated ≥5 bills · receipts per member

Every measure carries its n. Below minimum sample size, we leave it blank rather than guess. Embeddings: OpenAI text-embedding-3-small (1,536-d). Refreshed weekly; every score links back to the cited statements behind it.

Read the full methodology — how Perch measures behavior →

Who it’s for

For the people who read the room for a living.

If your job is knowing where a member stands before you walk in, Perch is the prep.

Lobbyists
Walk into every meeting knowing where they stand — and what actually moves them.
Advocacy & nonprofit leads
Find who’s persuadable on your issue before you spend a relationship finding out.
Government affairs teams
Member intelligence in minutes — the read you’d get from years in the building.

Ready to know where they stand?

Request access to Perch

Already have access? Sign in →

Security & privacy

Your research stays yours.

A Perch search can reveal who you’re working for. So it’s built to keep that to yourself — encrypted, private, and never used to train a model. Privacy Policy →

Encrypted in transit Encrypted at rest Access restricted Stripe billing No model training Every claim cited

Confidential by design

Your searches, briefs, and history aren’t visible to other subscribers, to the members you research, or to anyone at Perch. Yours alone.

Encrypted end to end

Encrypted in transit and at rest. Production access is locked down, and billing runs through Stripe — we never touch your card number.

Never trains on your work

Your queries and briefs are never used to train AI models — not ours, not anyone’s. Your strategy doesn’t become someone else’s feature.

Everything’s checkable

Every claim links to the statement it came from. AI can err — the [s1] citation is how you catch it, in one click against the record.