This cafe uses AI to monitor its staff efficiency and the length of stay of customers. 🫣 Thoughts?
Today:
- User insights you can build on
- Never have I ever won this game
- How to run LLMs locally
Feedback from millions in minutes
Since Kraftful’s shift from the IoT space in February, founder/CEO Yana Welinder has been making waves. She’s been featured on Reid Hoffman’s ‘Masters of Scale’ and in a Fast Company roundup of “San Francisco cool kids racing to build AI’s next big thing.”
In addition, product teams from tech giants like Google, Microsoft, Meta, Netflix, Canva, Dropbox, Atlassian, and LinkedIn have all adopted Kraftful. Following their rigorous testing, Kraftful’s latest iteration launches today.
Copilot for product builders: It’s a tool that connects to multiple qualitative feedback sources — think Zendesk, Intercom, Delighted — to aggregate input from your clients into actionable, AI-generated summaries, sent daily, weekly, or monthly, to your inbox or Slack. It also creates AI-written user stories complete with acceptance criteria, which can be synced to Jira, Linear, Trello, or wherever dev work gets done.
Kraftful, which raised 1M back in 2020, originally aimed to help smart home companies make better apps. The company counts Google Assistant Investments Program among their backers, along with YC, F7 Ventures, Cleo Capital, and a number of other angel investors.
A GPT-enabled innovation spillover: “Pivoting was hard but the time (and tech) was right,” says ‘One Knight in Product’ podcast host Jason Knight, who interviewed Welinder pre- and post her pivot from IoT to AI-powered product development. “Yana was solving a problem she deeply cared about, but when ChatGPT hit primetime she realized she could solve an even more important problem.”
Your agents stopped answering and started doing
Your agents call tools, write to databases and move money, so a wrong answer is now a wrong action, which is a considerably worse meeting. Arcjet runs policy before every LLM call, tool call, query and API request: prompt injection, sensitive data, bots, rate limits, spend quotas, and your own rules in Rego on top. It reads the OpenTelemetry you already emit, so switching it on is one environment variable, no code changes and nothing to deploy. Each check also reads the earlier steps in the same workflow, so an agent's tenth action gets judged against the first nine and not on its own. JavaScript, Python and Go SDKs all at 1.0, agent framework guards included, free tier to start on.
Daily Top Products
- Gridea is a blogging platform with a ‘Thought Box’ for research, as well as hosting and custom domain features.
- Weather forecast by Rainbow AI uses machine learning to provide more accurate weather predictions.
- SEC Insights AI adds a chat interface to help users analyze SEC documents in natural language.
- Never Have I Ever is a party game app with 800+ cards and AI-personalized challenges.
- morningmate makes project management look and feel as simple as using social media.
- RoboResponseAI is a no-code chatbot trained to resolve customer queries before they become tickets.
- Ollama lets makers run Llama 2 and other large language models locally.