You can add a working AI assistant to any web app in a few minutes, without writing a line of code. In this video I do it inside a random product (Jira, but it could be any tool you already log into). Then I ask it to find my onboarding tickets and assign one to Mark. It just... did it. Why bother? Building an AI assistant that actually works inside your product usually takes weeks of engineering. You have to define every action the AI can take, keep its knowledge in sync with your UI, and rebuild pieces of it every time you ship a change. Most teams either skip it, or bolt on a chatbot that only knows the docs and can't actually do anything. This demo shows a way to skip that whole loop. This is Frigade Skills. An AI agent that learns your product by using it, then handles user questions and takes actions on their behalf. Think of it as an AI teammate that lives inside any web app you already log into. Jira, Spotify, Hacker News, whatever. What happens in the demo: - I invite the Frigade agent to Jira. It logs in and explores the product. - The Frigade dashboard fills up with discovered skills. Create issue, search issues, assign issue. - I turn on a few skills, then go back to Jira and ask the assistant to do things. - It handles ambiguity. When I said "assign to Mark" and there were multiple Marks, it asked which one. If you ship a SaaS product, this replaces the manual work of building user onboarding flows and authoring product tours. The agent learns your product itself and re-learns it when you change things. Same idea as a digital adoption platform, but nobody has to author every flow. Try Frigade → More demos (Spotify, Hacker News, Jira) → #AI #NoCode #AIAgent #SaaS #UserOnboarding #DigitalAdoption #InAppAI Frigade: YC Winter 2023: Frigade Skills: the in-app AI agent that learns your product and acts for your users
