The Visibility Paradox: Why AI Infrastructure Companies Can't Be Found by AI
DevTune.ai tracked AI search visibility across 13 companies in the agent authentication space. The category leader appears in AI responses 27% of the time. Four companies are invisible. Here is what it means.
Half of B2B software buyers now start their product search inside an AI chatbot, not Google. According to G2's 2026 buyer research, the share of buyers who begin with an AI assistant grew significantly over a four-month measurement window.
Which raises a question the AI infrastructure space doesn't seem to be asking: when a developer asks ChatGPT or Claude "what's the best platform for agent authentication?", does your company show up?
For most: no.
The data
DevTune.ai tracks AI search visibility across companies in the agent authentication vertical — one of the fastest-growing B2B SaaS categories in 2026 (projected $13.5B by 2032, growing 42% annually).
The results are sobering:
| Company | AI Visibility | |
|---|---|---|
| Auth0 (Okta) — category leader | 27.2% | |
| WorkOS | 26.4% | |
| Nango | 20.0% | |
| Stytch | 19.2% | |
| Composio | 17.6% | |
| Merge | 15.2% | |
| Arcade | 8.8% | |
| Descope | 5.6% | |
| 4 companies | 0% — invisible | |
Source: DevTune.ai AI visibility rankings, agent authentication vertical, June 2026.
Think about what this means. The companies building authentication infrastructure for AI agents — whose entire product thesis is that AI needs better tooling — cannot themselves be found by AI.
Why this matters more than you think
When a developer Googles "agent authentication platform", they see ads, organic results, comparison pages. They click through, evaluate, decide. This is familiar territory.
When that same developer asks Claude or ChatGPT the same question, the AI gives a direct answer. Maybe it names three options. Maybe five. If you're not in that answer, you don't exist. There's no page 2. There's no "next" button. The AI gave its recommendation and the developer moved on.
This isn't a vanity metric. It's pipeline. Every query where you're absent is a decision made without you on the shortlist.
Why the gap exists
Most B2B SaaS companies optimized for Google's algorithm over the past decade. They built keyword-targeted landing pages, wrote comparison posts, invested in backlinks and domain authority.
None of that is what AI models use to form recommendations.
AI models synthesize from structured information: documentation quality, community discussion depth, how clearly your product's unique value is articulated across trusted sources, how consistently your positioning shows up in technical forums and developer communities.
Companies like Composio and Nango have started playing this game — both publish "best agent authentication platform" comparison posts clearly designed to be picked up by AI. But their visibility scores (17.6% and 20%) suggest they're treating AEO as content marketing, not as a systematic optimization discipline.
Three things that actually move AI visibility
1. Structured, authoritative product documentation
AI models give disproportionate weight to well-structured docs. Not marketing pages — actual technical documentation that clearly states what your product does, how it differs, and what problems it solves. Vague docs produce vague AI recommendations.
2. Third-party technical validation
Blog posts on your own site help, but AI models weight independent sources heavily. Developer community discussions (Stack Overflow, HN, Reddit), integration partners' documentation, and independent reviews signal genuine market presence vs. marketing noise.
3. Consistent positioning across surfaces
If your landing page says "agent-first authentication" but your GitHub README says "OAuth management for APIs" and your Dev.to post says "integration platform", AI models can't form a clear recommendation. Inconsistent positioning gets averaged into irrelevance.
Who should worry
This isn't just an agent authentication problem. Every B2B SaaS company selling to developers is exposed to this shift. Developer tools, infrastructure, DevOps, security — any category where buyers research inside AI.
The companies most at risk:
- Strong product, weak documentation — AI can't recommend what it can't understand
- Small but growing — the AI recommendation gap compounds. Leaders get recommended, get more users, generate more community discussion, get recommended more
- Category creators — if you're defining a new category, AI models haven't seen enough context to recommend you. You need to deliberately build that context
The compounding dynamic is the most dangerous part. Auth0 at 27% doesn't stay at 27% — their visibility generates more community mentions, which increases their visibility further. The gap between the category leader and the zero-visibility companies will widen, not close, without deliberate intervention.
How does your product score in AI search?
Same engine our paid analysis runs on. Score, gap list, and inline card checkout — all in English. Free, runs live, no email required.
Score your site for free →