Built This Week/Guests

Guests

Matias Singer

Matias Singer, a Technical Program Manager at Ryz Labs, built RADAR with Claude and Codex two months after he started building. It turns meeting transcripts into Jira tickets.

Matias Singer is a Technical Program Manager at Ryz Labs, Jordan and Sam's company, where he works on EntreVista. He is not a developer. He had been on the team for five months and had only started building with Claude and OpenAI Codex two months earlier when he joined us in Episode 39 (April 17, 2026). He came on to show RADAR, the internal app he built to take the busywork out of his job.

It reduced my time that were four hours to minutes.

— Matias Singer, Episode 39

Why he built RADAR

Matias began by asking developers about their repetitive work and how many tools they juggled. His own pain was standups. If he was presenting, he couldn't type the updates. Not being a developer, he sometimes misheard what engineers meant. When a ticket wasn't raised in time, work got delayed.

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

The demo

RADAR is open to everyone at the company, with sections for boards, analytics, transcripts, weekly business reviews (WBR) and tail tickets. At 03:31 Matias showed boards for internal projects across the social, marketing and IT teams, mirroring their Jira boards, and an analytics view with charts. He added charts because early meetings spent staring at a doc were boring.

The main feature came at 04:42. Matias pasted a standup transcript, recorded with Granola and in Spanish, and RADAR suggested five tickets with assignees from the project's members. As Sam pointed out, nothing goes to Jira automatically: Matias can edit each one, deselect what isn't really a ticket, and then create the rest.

At 09:07 he showed the WBR section. You upload one or more transcripts and RADAR writes an editable weekly report per project with key matters and upcoming priorities. Matias said it sometimes surfaces things that aren't that important, which is why the report is editable.

Key ideas from the conversation

  • Four hours down to minutes (06:45). Ticket work that used to eat his day now takes a paste and a review.
  • Accurate, but keep a human in the loop (07:18). Matias called it pretty accurate and said people still need to oversee it.
  • The roadmap is less pasting (07:48). Next up are tail tickets in WBRs, a pre-meeting section, stored transcript history and a direct Granola integration so the meeting itself feeds RADAR.
  • Weekly show-and-tell drove adoption (10:40). Ryz Labs runs separate AI sessions for technical and non-technical staff, led by a senior developer, where people show what they built.
People that never work with a developer, the sales team, marketing, finance team, they are all working on their own apps.

— Matias Singer, Episode 39

How the rest of the team builds

Sam described the path that got people like Matias building: demo sessions that walked through basic builds, access to the tools, then a weekly meeting where someone almost always presents something new. The accounting team built cash flow and AP/AR dashboards, and the social and sales teams built their own tools, many hooked into Jira, Granola and Slack. Jordan noted it is still fairly technical work, with back ends, front ends and GitHub, but it keeps getting closer to building apps just by describing them. For more, see our AI in productivity page.

FAQ

Who is Matias Singer?

Matias Singer is a Technical Program Manager at Ryz Labs, Jordan and Sam's company, where he works on EntreVista. He is not a developer and started building with Claude and Codex about two months before Episode 39.

What is RADAR?

RADAR is an internal Ryz Labs app Matias built. It shows Jira-style boards and analytics across internal projects, turns pasted meeting transcripts into suggested Jira tickets, and generates editable weekly business review reports.

Can AI turn meeting transcripts into Jira tickets?

Yes. In RADAR you paste a standup transcript, even one in Spanish, and it suggests tickets with assignees. You edit and select which ones to create, so a person still reviews everything before it reaches Jira.

How much time does RADAR save?

Matias said it cut work that used to take him about four hours down to minutes. He called it pretty accurate but said it still needs human oversight.