Where you get picked
Recommendation rate by answer surface, with the same prompt set and scoring rubric.
The prompts that matter
A result is a recommendation, shortlist appearance, or absence—not a keyword rank dressed up as one.
Who agents choose instead
Why you’re losing
Evidence gaps found on the public web—not invented access to a model’s private reasoning.
Your “easy setup” claim has no retrievable proof
Agents repeat setup-time evidence for Help Scout and Plain. Example says “get started quickly,” but offers no number, customer quote, or walkthrough an answer can cite.
You do not own a specific customer or constraint
Winning products are described with a sharp fit—“for bootstrapped teams,” “for live chat,” or “for complex support.” Example is summarized as a general support platform.
The pages agents cite do not mention you
Six independent comparison and directory pages account for most non-vendor citations. You appear on one; the category leaders appear on five or more.
What to ship next
Ranked by likely recommendation impact, grounded in the gaps above.
Publish an evidence-led Intercom alternative page
Lead with the no-sales-call constraint, show exact price at 5 and 20 seats, and substantiate setup time with a measured onboarding example.
/intercom-alternative-for-startupsReplace “easy setup” with a verifiable claim
Add median time to first reply, the sample size, and a customer quote beside the onboarding CTA on pricing and comparison pages.
/pricing + 3 comparison pagesEarn inclusion in three cited category sources
Prioritize the independent pages already retrieved in lost prompts. Pitch a factual product brief, not a backlink exchange.
3 third-party sources