AI testing startup Momentic raises $15M to speed up automated software quality assurance with agentic AI.
Testing — which some users refer to as think time, that is the amount of user time needed in order to understand and validate something or execute a task — has taken on a lot of importance in recent years, especially among developers who tinker with bots and other services that underpin apps and platforms.
Standard Capital led the round, with contribution from Dropbox Ventures, following a $3.7 million seed earlier this year.
Momentic’s funding round and product vision explained
Momentic’s proposition is relatively simple: do away with fragile, handwritten test scripts in favor of lines authored by AI that understand products the way their users do. Wei-Wei Wu and Jeff An, cofounders who previously worked in developer tooling at companies like Qualtrics and WeWork, say the teams should validate features by describing important paths — search, sign-up, checkout — through natural language. It’s also a CI system that writes its own tests. The tests run everywhere, and you never have to fight with selectors or brittle edge-case logic.
The company says it now has 2,600 users and customers include Notion, Xero, Bilt, Webflow, and Retool. By its own measures, the platform performed 200 million automated test steps in the past month, which indicates that test volume will scale up rapidly when authoring and maintenance friction goes down.
How Momentic’s AI-driven testing platform works today
Momentic is built on established execution layers (browsers, mobile runtimes), like frameworks such as Playwright and Selenium, but it replaces manual scripting with so-called task-oriented agents. Teams write goals in English, append some guardrails (e.g., what specific data is available or certain expected states on the page), and we produce step-by-step actions with assertions. With each UI change, a “self-healing” layer adjusts locators and updates flows instead of breaking tests left and right.
Under the hood, the authoring experience focuses on maintaining and restoring state. The agent can retry with context if login fails due to a transient network error instead of giving a false negative. At scale, in CI pipelines where thousands of tests run concurrently, it matters. Mobile app support was added by Momentic in August, and deeper test case management will be implemented so teams can manage suites for web and mobile together with centralized ownership, history, and flaky-test triage.
Competitive landscape for AI-powered software testing tools
Open-source incumbents like Playwright and Selenium are still the default for engineers who need complete control and low-level hooks. The issue is whether AI can approximate the same accuracy in a world of significantly reduced setup and maintenance costs. Meanwhile, the foundation model providers themselves — chiefly OpenAI and Anthropic — are releasing patterns for agentic testing, opening up the possibility that companies could cobble together their own AI-driven harnesses.
What is Momentic’s response? Invest in our strengths: reliability, enterprise ergonomics, and lifecycle features — governance, auditability, CI/CD integrations, and analytics that bring to light flakiness and risky code paths. If it can reliably cut down on false positives, clamp selector turnover in the face of UI churn, and speed up feedback loops, then it will clear a defensible space for itself — even though model capabilities evolve, you still have to keep using reliable tooling even after they’ve improved.
Why AI-driven software testing matters for modern teams today
Today’s teams ship faster than your traditional QA org could ever press a button for. DORA research has long associated high deployment frequency and low change-failure rates with powerful automation, including testing — yet many organizations struggle with brittle “tests-all-the-way-down” test suites that slow down releases or give false alarms. Side by side, low-code tools and AI-assisted coding are proliferating in production — each with user journeys that must be continuously verified.
Industry analyses from organizations such as Gartner and Forrester have identified AI-augmented quality engineering as a major trend, emphasizing that bugs found early in the lifecycle are orders of magnitude cheaper to fix than those unearthed in production. The World Quality Report also cites the move to continuous, risk-based testing instead of manual validation. In that context, tools that translate product intent straight into executable tests can shorten the cycle time and increase coverage across browsers, devices, and locales.
For example, an e-commerce team might define a full checkout flow — which includes such things as promotions, address validation, and multi-currency payments — and have the tests scheduled to run on every pull request on Chrome, Safari, and mobile. Should the payments UI change, self-healing logic updates selectors and maintains assertions, reducing the maintenance burden that so often leaks into sprint capacity.
Execution risks, reliability challenges, and what comes next
For any AI-first testing vendor, the single biggest risk is brittleness in the guise of intelligence. The agents that overfit happy paths, silently miss failing assertions, or produce flaky flows can erode confidence. That will drive Momentic to deploy the fresh capital toward reproducibility and sandboxed actions, with transparent diffs between generated and approved test logic slated for development, as well as deeper controls for security and data handling inside regulated environments.
With its new round of funding, the company intends to grow engineering; refine mobile features; and develop more advanced test case management. If it can continue to scale reliable execution — avoiding flaky tests even as the range of test cases grows — Momentic will be well-placed when enterprises start looking to combine AI-assisted coding with AI-powered assurance. The more code, as Wu says, the more trips to verify. The teams that will ship the fastest are the ones who can trust their automation.