GC AI sells to lawyers, the most careful readers there are, and its proof points disagree with each other. The Harvey comparison says NPS 77, the LegalOn comparison and the time-to-value page say 74, and the job posts say 80. All 13 comparison pages say "200+ public companies and 25 unicorns" while the job posts say "300+ public companies, 75+ unicorns". Lawyers trained is 6,000+ on the comparison pages and 8,000+ on the About page; the banner counts 12M+ contracts and the About page 9M+ documents. The llms-full.txt file that AI assistants are invited to read repeats the mix. The new Director of Product Marketing is asked to keep the website and sales materials consistent and grounded in evidence.
First action
List every figure on the 13 comparison pages, the About page and the time-to-value page in one sheet.
How to measure it
Pages with current, dated proof figures
- Now
- –
- Target
- 100%
- By
- 30 days
Steps
Inventory the figures
Every number on comparison, About, Academy, banner, press and job-post pages, and in llms-full.txt.
Director of Product Marketing 2 days
Pick the current value
One agreed number, date and method for each claim.
Director of Product Marketing 2 days
Check the method
NPS survey size and date, customer-count definition, what counts as a unicorn.
Head of Revenue Operations 2 days
Build one source
A shared data file or CMS field that every page reads.
Member of Technical Staff, Growth Engineering 1 week
Replace hard-coded numbers
Comparison pages first, then About, Academy and the banner.
Senior Content Marketing Manager 1 week
Regenerate llms-full.txt
From the same source, so assistants get the same figures.
SEO and AEO lead 1 day
Align sales materials
Decks, competitor comparisons and email templates use the same sheet.
Director of Product Marketing 1 week
Review monthly
Update figures on a set day, with the date shown.
Director of Product Marketing monthly
Why it works
Create one dated proof sheet (NPS with survey date and sample, customers, public companies, unicorns, lawyers trained, contracts reviewed, time saved, savings) owned by product marketing. Render every figure on the site from that one source, so the comparison pages, About page, banner, Academy, job-post boilerplate and llms-full.txt all update together. Put the date and method next to each figure ("NPS 80, customer survey, September 2026"), and retire numbers that are no longer current.
A skeptical general counsel who sees 74 on one page and 80 on another starts doubting every other claim, and so does a procurement reviewer. One source also means AI assistants quote the newest and strongest numbers, which today are hidden in the job posts.
Ready to use
ChecklistProof sheet fields
[ ] Customers (definition, date) [ ] Public companies and unicorns (definition, date) [ ] NPS (survey date, sample size) [ ] Lawyers trained (through Academy, date) [ ] Contracts or documents reviewed (which one) [ ] Time saved and savings (study, date, sample) [ ] Countries [ ] Where each figure appears (pages, llms-full.txt, decks, job posts)
Risks and how to handle them
Older, lower figures were accurate when published
Keep them in the study pages with their dates; use only the current figure in claims.
Bigger numbers invite scrutiny
Link each figure to its method so it can be checked.
Evidence
- Your websiteProof figures on gc.ai, checked October 22, 2026: the GC AI vs Harvey comparison says "NPS 77 (published, comparable to Apple and Netflix)"; the GC AI vs LegalOn comparison says "NPS 74"; the time-to-value page says "GC AI's Net Promoter Score: 74"; the job posts say "80 NPS". All 13 comparison pages say "200+ public companies and 25 unicorns", against "300+ public companies, 75+ unicorns" in the job posts. The comparison pages say "6,000+ lawyers trained"; the About page says "8,000+ Legal professionals trained" and the CEO's profile "More than 8,000 lawyers ... as of August 2026". The site banner says "12M+ contracts reviewed"; the About page says "9M+ Documents and workflows accelerated".Source
- Your websitegc.ai/llms.txt invites AI agents to read "the full corpus" at /llms-full.txt. That corpus repeats the mixed figures: "200+ public companies and 25 unicorns", "300+ public companies and 25+ unicorns" and "300+ public companies" in different passages, all dated September or October 2026.Source
- HiringEvery GC AI job post: "More than 2,100 companies run their legal work on GC AI, including 300+ public companies, 75+ unicorns"; "The platform holds an 80 NPS"; "GC AI has doubled ARR in six months"; a $60M Series B led by Scale Venture Partners and Northzone in November 2025; the team is across North America "and soon Europe".Source
- HiringDirector of Product Marketing (posted September 24, 2026): GC AI "is expanding across workflows, pricing tiers, and customer segments"; the role owns "narrative, positioning, pricing, packaging, launches, enablement, and competitive strategy", must "Maintain consistency across the website, sales materials, product surfaces, and customer education", oversee "case studies, return-on-investment stories, comparisons, and proof points" and "Ensure customer-facing content is accurate, useful, and grounded in evidence", and run win-loss research.Source