Autosana

What do they actually do

Autosana runs a cloud service that turns plain‑English test descriptions into end‑to‑end mobile tests on iOS and Android. Teams write “flows” in natural language, and an AI agent interprets and executes them on virtual devices, regardless of framework (native, React Native, Flutter) YC page docs: Flows docs: Introduction.

In practice, teams upload a build or connect CI so new builds are tested automatically, then run flows and review step‑by‑step screenshots or replays. The product includes self‑healing behavior to reduce brittle test failures, organizes flows into suites, supports automations and notifications/hooks, and provides CI/CD examples for GitHub Actions docs: Quickstart docs: Settings docs: Automations/Suites docs: CI/CD.

Evidence of production use includes a public pilot where Lucra reported automating about 70% of a 250‑step regression suite in under 15 hours, citing less manual maintenance due to the self‑healing approach Reuters. The founders frame the product as a “24/7 AI QA engineer” that runs on every build via CI YC page.

Who are their target customer(s)

• Mobile engineering teams at fast‑shipping startups: They lose time running and fixing manual tests between frequent builds and need quick, automated feedback on each build to catch regressions early.
• QA engineers maintaining large, brittle test suites: They spend cycles rewriting selectors and chasing flaky failures instead of finding real bugs; they need lower‑maintenance tests and clearer failure context.
• Small product teams without dedicated QA: They can’t afford full manual regression passes and often ship user‑visible bugs; they need a simple way to author and run tests without specialized QA staffing.
• Release managers / product owners: They need reliable pre‑release signals and failures routed to their existing triage tools so they can make ship/no‑ship decisions quickly.
• Mid‑to‑large enterprises with multiple apps/devices: They struggle with scale, onboarding many apps, device/OS coverage, and integrating testing into enterprise toolchains with auditability and access controls.

How would they acquire their first 10, 50, and 100 customers

• First 10: Founder‑led pilots with YC/startup network apps: personally onboard builds, author a few natural‑language flows, iterate to reliable runs, and use the Lucra pilot as a proof point to shorten cycles docs Reuters.
• First 50: Product‑led growth via CI examples and tutorials: publish GitHub Actions templates and short guides so teams can self‑trial and see value within 1–2 builds; support with targeted webinars/outreach to mobile teams docs docs.
• First 100: Partnerships and focused enterprise POCs: work with mobile dev agencies/CI vendors/test consultancies; run paid 2–4 week POCs that integrate notifications/hooks and release gates, then templatize onboarding playbooks docs docs.

What is the rough total addressable market

Top-down context:

Core market: mobile application testing services is estimated around USD ~6.6B in 2024 and growing at a mid‑to‑high‑teens CAGR Testlio summary. Adjacent market: broader automation testing (mobile+web+desktop) is ~USD 17.7B in 2024 with strong growth Fortune Business Insights.

Bottom-up calculation:

Assume ~200,000 companies worldwide actively ship mobile apps, each spending ~USD $30k/year on mobile testing tools/services; that implies roughly $6B TAM, broadly consistent with top‑down estimates. The scale is plausible given the global developer base (~19.6M in 2024) and prevalence of mobile initiatives JetBrains.

Assumptions:

• ~200k companies maintain actively shipped mobile apps.
• Average annual mobile testing spend ≈ $30k per company (tools + services).
• A meaningful share of the ~19.6M global developers work on mobile apps.

Who are some of their notable competitors

• BrowserStack (App Automate): Real‑device cloud to run existing automated tests (Appium/Espresso/XCUITest) with CI/CD integrations and logs. Competes on device coverage and infra; expects scripted tests rather than natural‑language agents.
• Sauce Labs (Real Device Cloud): Enterprise device cloud and test analytics with AI‑led insights. Overlaps on CI execution and diagnostics but centers on running/analyzing scripted suites, not converting plain language into tests.
• Firebase Test Lab (Google): Cloud device lab for instrumented tests and “Robo” exploratory checks integrated with Firebase/Play. Useful infra and pre‑launch signals; not an NLP/AI authoring agent for end‑to‑end flows.
• Applitools: Visual AI for detecting regressions and reducing maintenance. Primarily a visual/assertion layer added to existing tests rather than a stand‑alone agent that executes full mobile flows from natural language.
• Kobiton: Mobile‑focused testing with scriptless automation, Appium self‑healing, and real‑device cloud. Closer functionally via no‑code/recorded flows, but differs from Autosana’s free‑form natural‑language agent model.