LRG LABS
AI market intelligenceIndependent research · 2026

Know what AI
tells your market.

Buyers are asking AI about your category, your product, and your competition. We make those answers visible, measurable, and useful.

Explore the live proof of concept See what we measure
28claims captured
64%not fully correct
4AI assistants measured
0 / 458claims gave date + source
01 / THE SHIFT

AI is now a market surface.

Search rankings show who gets found. AI answers reveal who gets considered, how products are described, and which facts buyers carry into the room.

LRG LABS builds evidence-led research systems for this new layer of the market. We test real buyer questions across models, verify the claims against primary sources, and show vendors exactly where perception and reality diverge.

02 / WHAT WE MEASURE

From opaque answers
to decision-grade signals.

Not sentiment. Not synthetic personas. Observable claims, tested repeatedly and anchored to evidence.

A

Market presence

Who appears when buyers ask for a shortlist—and who vanishes when the model starts researching.

01Zendesk20/20
05Freshdesk14/20
08Front8/20
B

Claim integrity

Every product, pricing, and compliance claim is decomposed and judged against a verified record.

64%didn’t hold up
C

Information decay

When prices, capabilities, or policies change, we trace which models keep repeating the old reality.

JUL ’25$79 JAN ’26$89 AUG ’26AI: $79
D

Competitive substitution

See which alternatives enter the answer, what AI says about them, and whether those comparisons are defensible.

“The assistants shaping your buyers’ shortlists misprice competitors’ add-ons, invent plan gates, and cite pages that say no such thing.”
Open the competitive view
03 / CURRENT PROOF OF CONCEPT

The answer layer,
made inspectable.

Our helpdesk pilot turns hundreds of model outputs into an audit that a product, marketing, or competitive intelligence team can actually act on.

64%of measured Freshdesk claim-parts did not fully hold up against published evidence.
01

An invented SCIM capability traced to a third-party page—and a dead citation.

FALSE / UNSUPPORTED
02

A six-month-old price repeated even when assistants searched the live web.

STALE / LOCALISATION
03

A vendor recommended in 20/20 memory-only runs, then absent once research began.

ZERO-ROW
Explore the Freshdesk audit
04 / METHOD

Research you can
trace all the way down.

Every conclusion connects back to the answer, prompt, source, capture date, and immutable evidence hash.

  1. 01

    Frame

    Freeze real buyer questions and define the market panel.

  2. 02

    Observe

    Run repeatable captures across models, tiers, and search modes.

  3. 03

    Verify

    Decompose claims and judge them against first-party evidence.

  4. 04

    Monitor

    Track drift, decay, and competitive movement over time.

LRG LABS / AI MARKET INTELLIGENCE

See what the
market sees.

We’re working with software vendors who want an evidence-led view of how AI represents their market.

Explore the proof of concept