On Built This Week, AI productivity mostly looks like people fixing their own busywork. Sam replaced a stand-up form with an AI summary in about two hours. Matias Singer, a program manager on our team, built an app that turns meeting transcripts into Jira tickets. Our guests took the same idea further: Jennifer Dulski of Rising Team built an AI coach for managers, Adir Ben-Yehuda of Autonomy AI says his customers open fewer Jira tasks, and Hooman Radfar of Collective is moving a solopreneur back office onto agents.
So it was so annoying because if you were presenting, you were not able to type down what were the updates or a lot of times you miss things.
— Matias Singer, Episode 39
How companies we talked to use AI in productivity
Rising Team: an AI trainer for managers
Jen calls Rising Team's coach, RT, a personal trainer for leadership. It runs an intake, writes a one-of-one growth plan, books sessions on your calendar and lets you rehearse a hard conversation with a specific teammate. It knows your team from Rising Team's team-building sessions, which is the difference from a general chatbot. Sessions average 8.5 out of 10, and she cited a 10 to 33% lift in retention. (Episode 52)
And the BCG data says 70% of whether your AI transformation is successful or not lands on whether the people are able to adapt to these tools and still work with each other.
— Jennifer Dulski, Episode 52
Autonomy AI: less time in Jira and Figma
Autonomy AI lets product managers and designers change an existing codebase and send a pull request. The surprise for Adir was what customers stopped doing: they open fewer tasks in Jira, use less Figma, and "just need to act on the task." (Episode 43)
Collective: agents in the back office
Collective runs company formation, bookkeeping and taxes for solopreneurs. Hooman said the team benchmarks each internal flow for time and cost, then moves work like account categorization to an agent-based approach, in rank order, for faster and more consistent results. Members can already ask questions about their accounts, and his goal is a "CFO" agent in your pocket. (Episode 14)
What we built
- Daily Check-in: each team logs what it finished, today's priorities and blockers. Every morning an AI summary, blockers first, posts to Slack. Sam built it in about two hours between meetings with Bolt, Supabase and a GitHub cron job. The day before we recorded, it flagged that the sales team had run out of credits for a tool, and Sam fixed it right away.
- RADAR: Matias pastes a stand-up transcript, even one in Spanish, and gets suggested Jira tickets with assignees that he edits and selects. It also writes weekly business review reports. It cut about four hours of work to minutes and spread from his own use to teams across the company.
- Mate: Jordan's company directory for a mostly remote team, with profiles, tenure, interests and clubs. He built it over about a weekend with Bolt, Cursor and Supabase.
- Solopreneur finance dashboard for Collective: an AI cash-flow forecast, insurance check and lending-readiness report from an uploaded P&L. Jordan moved to Gemini for the big documents and admitted the overall AI insights panel may not work that well.
RADAR came out of how Ryz Labs, Jordan and Sam's company, got non-technical staff building. Sam described demo sessions, tool access and a weekly show-and-tell, with separate tracks for technical and non-technical people. The accounting team has built cash-flow and AP/AR dashboards, and sales and social teams have built their own tools.
What's still hard
- A person still checks the output. Matias said RADAR is pretty accurate but needs oversight, and Sam liked that it doesn't push tickets to Jira automatically.
- Prototype to production. Hooman's own team underestimates how quickly a prototype comes together now, he said, and getting from there to production is "the big last mile."
- Permissions. Daily Check-in had none at launch, so Sam asked teams to keep sensitive details out.
- Still technical. Even with AI, a tool like RADAR has a database, a server side and a code repository behind it. Jordan said that is still a lot to ask of someone who has never coded.
- Memory: helpful or creepy? Jen asked us directly. Jordan voted helpful: he wants an assistant that notices he forgot to set tomorrow's alarm.
Distribution matters too. In Episode 5 we covered Google's report of 50 million people using AI meeting notes in Google Meet. Sam said we're both pretty AI-forward, and he wasn't sure he had ever used it.
Episodes on AI in productivity
- Episode 5: Meet Mate, our internal directory (July 2025).
- Episode 7: How Daily Check-Ins Fix Remote Work (August 2025).
- Episode 14 with Hooman Radfar of Collective (September 2025).
- Episode 39: Meetings Into Jira Tickets with Matias Singer (April 2026).
- Episode 43 with Adir Ben-Yehuda of Autonomy AI (May 2026).
- Episode 52 with Jennifer Dulski of Rising Team (September 2026).
FAQ
What are good examples of AI productivity tools?
On our show: Sam's Daily Check-in, which summarizes team stand-ups with AI and posts blockers to Slack, and RADAR, which turns meeting transcripts into suggested Jira tickets. Guests showed Rising Team's AI leadership coach and Autonomy AI, which lets product people ship code changes.
Can non-developers build their own AI productivity tools?
Yes. Sam, who isn't a developer, built Daily Check-in in about two hours with Bolt, Supabase and a GitHub cron job. Matias Singer, a program manager, built RADAR with Claude and Codex after two months of building with AI.
How do you get a team to adopt AI tools?
At Ryz Labs we ran demo sessions, gave people access to the tools and started a weekly show-and-tell, with separate tracks for technical and non-technical staff. Sales, marketing, finance and accounting have since built their own apps.
Should AI tools act without human review?
Not yet in our experience. RADAR suggests Jira tickets but a person reviews and selects them, and Matias says it is accurate but still needs oversight. Rising Team's coach never helps with hiring, firing or rating decisions.