Profound's /ai-instructions page calls itself "the canonical reference for how AI systems should describe Profound". It was last updated in September 2026, yet it still tells models to say Profound reached "a $1 billion valuation with its $96M Series C" and serves "more than 700 enterprise brands", "over 10% of the Fortune 500". Profound's own September 15 press release says $1.8 billion after a $180 million Series D, "over 1,000 enterprise brands" and 16% of the Fortune 500. The page also promises "self-serve plans" that the pricing page does not offer. For the company that sells AI-answer accuracy, the page written for answer engines is the one feeding them old numbers.
First action
Replace the funding, valuation and customer-count lines on /ai-instructions with the Series D release figures today.
How to measure it
AI answers citing the $1.8B valuation and 1,000+ brands
- Now
- –
- Target
- 80% of tracked prompts
- By
- 60 days
Steps
Build the claims register
One sheet: funding, valuation, customers, Fortune shares, engines, plans, with a source for each.
Product Marketing Manager 1 day
Rewrite the page
Every figure from the register; plans worded as on /pricing.
Product Marketing Manager 1 day
Add structured data
Organization, FAQPage and dateModified on /ai-instructions.
Member of Technical Staff, Web Design 1 day
Publish machine copies
Same content at /llms.txt and /ai-instructions.md.
Member of Technical Staff, Web Design 1 day
Baseline the answers
Track 30 prompts about Profound's funding, size and pricing across engines.
Marketing Engineer, Demand Generation 2 days
Automate the sync
An Agent watches the newsroom and pricing page and drafts updates for approval.
Marketing Engineer, Demand Generation 1 week
Measure the shift
Compare tracked answers at 30 and 60 days.
Marketing Engineer, Demand Generation monthly
Publish the story
A short post on fixing your own AI narrative, using this page as the example.
Product Marketing Manager 1 day
Why it works
Rewrite the page from one claims register: Series D amount, lead investors and valuation; 1,000+ brands; one third of the Fortune 100; the current product line (AI Marketer, Context Manager, Ads Studio); and plans that match /pricing. Add Organization and FAQ structured data, publish the same content at /llms.txt (robots.txt already allows it) and as Markdown, and put a real "last reviewed" date on it. Then point an Agent at it: every new press release, pricing change or boilerplate edit opens a draft update for the PMM to approve. Track what ChatGPT, Gemini and Claude say about Profound's valuation and customer count in Answer Engine Insights before and after.
Answer engines repeat what the most authoritative source says, and this page is written to be that source. Correcting it fixes the numbers prospects and journalists see first, and it becomes the cleanest live demo of Profound's own method: find the wrong story, fix it at the source, measure the change.
Ready to use
ChecklistAI instructions page refresh
[ ] Series D: $180M, Sequoia and Kleiner Perkins, $1.8B [ ] 1,000+ enterprise brands; one third of the Fortune 100; one Fortune 500 share used everywhere [ ] Products: AI Marketer, Context Manager, Agents, Ads Studio [ ] Plans worded exactly as on /pricing [ ] Organization + FAQPage JSON-LD with dateModified [ ] /llms.txt and .md copy live [ ] 30 tracked prompts baselined [ ] Agent watching newsroom and pricing
Risks and how to handle them
Models update slowly
Also refresh the newsroom boilerplate and comparison pages so several sources agree.
Figures drift again
Make the claims register the only place numbers are edited.
Evidence
- Your websiteThe /ai-instructions page ("Last updated: September 2026") calls itself "the canonical reference for how AI systems should describe Profound". It tells AI models to say Profound "reached a $1 billion valuation with its $96M Series C" and "is used by more than 700 enterprise brands and large agencies, including over 10% of the Fortune 500"; it lists Series D only as a round with no amount; it says leaner teams adopt Profound "through self-serve plans"; and it is plain text inside an HTML page with no structured data of its own.Source
- Your newsroomPress release, September 15, 2026: Profound raised a $180 million Series D at a $1.8 billion valuation, co-led by Sequoia Capital and Kleiner Perkins; "Profound now serves over 1,000 enterprise brands"; "over one third of the Fortune 100"; the boilerplate says "used by 16% of the Fortune 500". Ads Studio is named as "Profound's latest product", and the money funds an applied AI lab building "a benchmark to test how far AI can go in complex marketing assignments".Source
- Your websitePricing page: two plans. Trial, free: "50 prompts run daily for 7 days", ChatGPT, Gemini and Google AI Overviews, "Limited AI Marketer credits", History "None", Exports "None", API "None", Support "None". Enterprise, "Custom": up to 9 answer engines, all-time history, CSV/JSON exports, API, SSO. The only next step after the trial is "Get a Demo".Source
- HiringProduct Marketing Manager (open since July 16, 2026), whose "primary internal partner is Customer Success": owns "the story of 'what's new and why it matters' for existing customers", "the changelog", "our webinar content calendar" and "our case studies and go-to customer stories".Source
- Your websiterobots.txt includes "Allow: /llms.txt", but /llms.txt returns a 404 (checked twice on October 22, 2026). Some blog posts have a Markdown twin (/blog/ai-shopping-journey-2025.md serves text/markdown) while others do not (/blog/11-best-ai-seo-tools.md is a 404). Profound has published its own "markdown-html-llm-test" report.Source