How Semaloop uses AI agents to automate mobile app testing
London startup Semaloop uses multimodal AI agents to test mobile apps on real devices in plain English, catching bugs before production.

Mobile app development teams face a persistent dilemma whenever they prepare a new release for production. Traditional end-to-end testing relies heavily on rigid code scripts and fragile accessibility identifiers that break whenever a screen layout or user flow changes. Because maintaining these brittle test suites consumes hours of engineering effort, many product teams end up skipping comprehensive end-to-end validation entirely. Consequently, developers lack confidence in their releases and frequently discover critical user-facing bugs only after an update reaches production.
To resolve this bottleneck in mobile quality assurance, London-based co-founders Charlie Kingston and Rory Smith created Semaloop. The startup builds multimodal artificial intelligence agents capable of testing mobile applications by directly mimicking real human interactions. Rather than writing long automation scripts or manually configuring test identifiers, engineers simply explain the flows, screen behaviours, and intended outcomes they care about using plain language. Semaloop then deploys AI agents that navigate real devices, tapping buttons, swiping through screens, and verifying visual and textual responses exactly as a human user would during an authentic session.
The platform is engineered specifically to handle the inherent visual and interactive complexity of modern mobile software. Semaloop agents evaluate applications across visual elements, text strings, and audio responses in real-world operating scenarios. As application user interfaces evolve, the agents adapt automatically, keeping tests useful without requiring constant manual rewrites from developers. When an agent discovers an anomaly, the platform surfaces complete contextual visibility into the run. High frame rate video replays show the exact sequence of interactions leading to the failure, helping engineers reproduce bugs rapidly.
Beyond identifying failures, Semaloop streamlines the remediation workflow by helping developers transition from bug detection to resolution with minimal friction. The system provides clear diagnostic context so teams can understand why an issue occurred and even assists in creating code fixes ready for merge. Having initially focused on mastering iOS application testing, Semaloop extended its agent technology to support Android devices as well. Development teams at mobile-first companies including Granola, Clove, and Ash by Slingshot AI now rely on Semaloop daily to protect their production releases.
To support its ongoing product development and team expansion, Semaloop closed a pre-seed funding round of over three million dollars. The investment was led by Concept Ventures, with additional participation from Evantic Capital, Weekend Fund, and FJ Labs. The startup also drew support from angel investors associated with leading tech firms like OpenAI, ElevenLabs, Hugging Face, Linear, Lovable, and Dropbox. As Charlie Kingston and Rory Smith prepare to showcase Semaloop at Slush 2026, the team remains committed to eliminating manual quality assurance friction for software developers globally.
Find out more at semaloop.com
The Future FrontBe on the cover.Twice a month we put the people shaping what's next on our cover.Apply for the cover





