Signals
AI agents are rewriting the go-to-market playbook
Buyers now research, shortlist and even book demos through AI assistants. Here is what that changes for startups trying to get found, and what to do about it this quarter.
The Third Orbit · · 2 min read
Something quiet happened to the top of the funnel this year. The founders we work with started hearing a new answer to "how did you find us?": "I asked an AI assistant for options."
It is not a fringe behaviour any more. Buyers are using AI assistants to do most of their early research: defining the problem, listing vendors, comparing pricing and drafting the shortlist. By the time they land on your website, the decision is half made.
What actually changed
Search engines ranked pages. Assistants summarise companies. That is a different job, and it rewards different things.
- Clarity over keywords. An assistant can only recommend you if it can explain, in one sentence, what you do and who you do it for.
- Specifics over slogans. "We help teams move faster" gives a model nothing to work with. "Payroll for remote teams in 40 countries" does.
- Evidence over claims. Case studies, numbers and named customers are what models quote when they justify a recommendation.
If an AI cannot describe your company in one sentence, it will describe a competitor instead.
The new top of funnel
For a startup between traction and scale, this is good news. Big brands win on budget. Assistants reward precision, and precision is cheap.
1. Make your positioning machine-readable
Your homepage headline, your About page and your product pages should all state the same three facts: what you do, who it is for, and what makes you different. Consistently, in plain language, on pages a crawler can read.
2. Publish answers, not announcements
Write the pages your buyers ask assistants about: comparisons, pricing logic, implementation guides, "how to choose" articles. Each one is a chance to be the source an assistant cites.
3. Give every claim a number
"Faster onboarding" becomes "onboarding in 9 days, down from 6 weeks". Numbers survive summarisation. Adjectives do not.
How to measure it
Most analytics tools still lump AI traffic into "direct" or "referral". Two simple fixes:
- Create a segment for referrers like chatgpt.com, perplexity.ai, claude.ai and gemini.google.com.
- Add one question to every demo form: "Where did you first hear about us?" Include "An AI assistant" as an option.
Within a quarter you will know whether assistants are sending you buyers, and which pages they are reading.
What to do this quarter
| Week | Action |
|---|---|
| 1 | Ask four assistants to describe your company and your category. Note the gaps. |
| 2 | Rewrite your homepage and About page to state what, who and why in plain language. |
| 3 | Publish one comparison or "how to choose" page your buyers actually need. |
| 4 | Add AI referral tracking and the "where did you hear about us" question. |
The companies that win this shift will not be the loudest. They will be the easiest to explain.