The Agent Default Report
What AI coding agents actually install, build with, and keep, measured from real agent sessions. These are the headline findings; each vendor's full funnel, switch graph, and failing runs live in its private scorecard.
Data through 2026 Q2 · last updated July 2026
Every category has a runaway default.
Ask hundreds of times, in different words, and one name keeps winning the pick. Selection is winner-take-most, set by what the model already knows.
| category | agent default | top-pick share |
|---|---|---|
| backend / BaaS | Supabase | 95% n=162 |
| feature flags | LaunchDarkly | 58% n=144 |
| transactional email | Resend | 54% n=162 |
| auth | Clerk | 51% n=243 |
| hosting | Vercel | 49% n=243 |
| voice AI | Vapi | 44% n=243 |
Being named isn't being picked.
The most common trap in the data: agents mention you constantly and choose you almost never. You're in the room, losing the close to one rival.
| SendGrid | named 82% → picked 15% |
| Firebase | named 46% → top pick 0% |
| Auth0 | named 33% → top pick 0% |
The switch graph shows who wins each vendor's losses. That part stays private.
Your standing depends on the model your customer runs.
The same tasks compile far more often on Opus than on Sonnet, and Sonnet is the model most customers run. A 22-point swing you don't get to choose.
The biggest competitor is often no vendor at all.
Building from scratch, agents hand-roll the integration 35% of the time (n=202, CI 29-42). DIY shows up in the recommendations too: an auth library like NextAuth or Lucia is named in 62% of auth answers, raw Nodemailer/SMTP in 42% of email answers, self-hosting in 19% of hosting answers.
Once you're in and building, you stay.
94% of installed, compiling integrations survive a "modernize this" refactor (n=806). Retention is rarely the leak. The whole fight is getting named, installed, and through the first build.
And the build stage is provably fixable: in controlled re-tests, a tailored in-repo fix doc took build success from 57% to 100% on Resend, and took a second email vendor from failing most agent builds to 96% (Sonnet; directional). The lift is biggest on the cheaper models your customers actually run.
Common questions.
Which tools do AI coding agents install by default?
In AgentRank's Q2 2026 controlled runs, every category had a runaway default: Supabase won 95% of backend picks (n=162), LaunchDarkly 58% of feature flags (n=144), Resend 54% of transactional email (n=162), Clerk 51% of auth (n=243), Vercel 49% of hosting (n=243), and Vapi 44% of voice AI (n=243). Directional figures.
Does being mentioned by an AI agent mean it will install you?
No. Naming and picking are different behaviors: in the same Q2 2026 runs, SendGrid was named in 82% of email answers but picked in 15%, and Firebase was named in 46% of backend answers with a 0% top-pick rate. Directional, controlled runs.
Do different AI models pick and build with different tools?
Yes, and the swing is large. The same from-scratch tasks compiled 87% of the time on Claude Opus (n=89, 95% CI 79-93) and 65% on Claude Sonnet (n=109, CI 56-73) in Q2 2026 controlled runs. Standing with one model does not transfer to another, so AgentRank reports every metric split by model.
How often do AI agents skip vendors and hand-roll the integration?
In 35% of from-scratch builds the agent wrote the integration itself instead of adopting any vendor (n=202, 95% CI 29-42, Q2 2026, directional). In recommendations, a DIY option like NextAuth or raw SMTP appears in up to 62% of answers depending on category.
Can a vendor change how AI agents build with their tool?
Yes, and it is measurable. In AgentRank's controlled re-tests, a tailored in-repo fix doc raised build success from 57% to 100% for Resend, and took a second email vendor from failing most agent builds to 96% (Sonnet, directional). The lift is biggest on the cheaper models most customers run.
Do AI agents keep the tools they install?
Mostly yes: 94% of installed, compiling integrations survived a modernize-this refactor pass in Q2 2026 runs (n=806, directional). Retention is rarely where a vendor loses; the fight is being named, getting installed, and surviving the first build.