Worktrace Ai

Worktrace AI develops an enterprise "AI adoption layer" that identifies which parts of a company's work can be automated with AI. A desktop application runs in the background on employees' computers and passively observes activity across applications, requiring no integrations. From that captured activity the software reconstructs workflows, turns them into visual process maps, and ranks candidate workflows by expected impact such as hours saved and alignment with company KPIs.

Beyond discovery, the platform produces recommendations for redesigning processes and generates blueprints for building the corresponding automations, which customers can implement on third-party agent frameworks including n8n, Crew AI and Asteroid. The company positions the product as an alternative to the conventional approach of manually authoring Standard Operating Procedures, interviews, shadowing and consulting engagements, which its founders describe as taking six months or more and producing documentation that is often incomplete or out of date. Worktrace states the product includes privacy controls such as PII redaction, allowlists and denylists, on-premise deployment and no productivity tracking, and references ISO/IEC 27001:2022, AICPA, GDPR and HIPAA compliance.

Worktrace markets to operations-heavy functions in financial services, logistics and transportation, insurance, healthcare, manufacturing, retail and CPG, and travel and tourism, citing use cases such as loan processing and underwriting, KYC/AML compliance, claims intake and adjudication, patient intake, order-to-cash and disruption recovery workflows.

Founding story

Angela Jiang was OpenAI's third product manager, working on the GPT models behind ChatGPT and later on public policy and AI literacy work before leaving in December 2024. Deepak Vasisht earned a PhD in computer science at MIT, worked at Microsoft including on the FarmBeats agriculture data project, and became an assistant professor of computer science at the University of Illinois. Both interned at Microsoft headquarters years earlier without meeting, and Jiang earned a PhD in deep learning from Carnegie Mellon before a product role at Determined AI (acquired by Hewlett Packard Enterprise) and joining OpenAI in 2021. They connected through overlapping AI circles and, while spending time at South Park Commons during Vasisht's leave from teaching, decided to build together. After exploring AI literacy and upskilling ideas, they concluded that instead of retraining workers on evolving AI tools, software should map their work and apply AI to it, which led to Worktrace.

Business model

Enterprise software sold to large organizations, with a desktop agent deployed across employee machines that generates workflow insights and automation blueprints; deployment options include on-premise. Automations are executed on third-party platforms of the customer's choosing rather than solely within Worktrace.

Not stated in sources beyond an enterprise sales motion; the site references a pricing page and a free trial offer.

Traction

At its December 2025 public launch the company had five design partners and a team of four full-time and three part-time employees, with the product developed in roughly eight months since founding. Published customer material describes a BukuWarung engagement covering high-ROI automations and a company-wide SOP repository, and another deployment giving visibility across roughly 20 teams. The site cites 90% of hours reclaimed across FTEs in ops-heavy teams and 96-99% accuracy validated with human-in-the-loop review.

Latest developments

Worktrace launched publicly in December 2025 alongside its $9.3 million seed round, and in April 2026 its co-founders were named to the Forbes 2026 AI 50 Brink List. Press coverage in April 2026 also noted OpenAI alumni investing from a new fund, with Zero Shot Fund describing Worktrace as its first investment.