šš½āāļø Stephenās Note
I think Iāve made my position clear on the matter ā performance reviews are broken. Six-month cycles make no sense in an AI-powered workplace where expectations change monthly.
Enter Thomas Forstner, VP of People & Talent at Juro. With a team of just four, he rewired performance management using AI. Instead of slogging through clunky feedback windows and retro reviews, Juro now runs monthly AI-guided feedback loops and future-forward role conversations.
This isnāt just a tool. Itās a playbook for eliminating PIPs and making performance management continuous, data-driven, and future-ready.
šŖš½ Running Lean with a Big Idea
Hereās the detail that stuck with me: Juroās People & Work Tech org is just four people.
Yet with that lean team, Thomas built a suite of custom GPTs and lightweight automations that many vendors would claim takes a year (and six figures) to roll out.
Itās not about scale. Itās about building with first principles.
ā The Pain
Like most companies, Juro relied on the ābest practiceā of six-monthly reviews. The result?
Managers couldnāt recall six months of detail.
Feedback landed too late to be useful.
One-hour retro + future planning sessions were at odds with the pace of AI-era change.
As Thomas put it: āBest practice is often just shorthand for not thinking from first principles.ā
For a lean, high-density team like Juroās, talent quality is everything. Old processes werenāt keeping up.
šļø Join the Ask Me Anything with Thomas Forstner
Thomas is ready to share how he rewired performance reviews with AI ā and why this could be the end of PIPs.
Bring your questions. Learn how he built it. See what continuous, AI-guided performance can mean for your team.
āļø The Build
Before writing a single prompt or setting up a Zap, Thomas and his team did something most HR leaders skip: they stepped back and reimagined performance management from first principles.
The starting point: what if reviews werenāt a compliance exercise at all, but a living system for conversations, decisions, and growth?
From there, they designed a suite of simple, agentic tools ā not platforms ā to make that system real:
OILSmith ā A custom GPT guiding managers through Juroās OILS framework (Observation, Impact, Listen, Suggestion). Instead of āwrite your review,ā managers prep feedback in minutes, practice tough conversations, and even role-play with the bot. The goal: make feedback a skill, not a task.
FutureSmith ā A GPT for āfuturespectives,ā shifting the conversation from what happened in the last six months to what roles might look like in six months ā especially in an AI-shifting workplace. This philosophical pivot is key: feedback isnāt just retroactive, itās future-proofing talent density.
Automation Layer ā Slack nudges remind managers when conversations are due. Humaans HRIS workflows and Zapier reporting handle the admin and pipe data back into the system. The result: performance as a continuous data stream, not a twice-yearly event.
Org Brain Backbone ā Notion serves as the living handbook and AI-integrated hub for everything: feedback logs, GPT templates, role documentation. Instead of static files, Juro has a knowledge system that learns as the org learns.
This wasnāt just āAI for HR.ā It was a philosophy turned into practice: small, modular agents solving one problem well, connected into a bigger system.
š§° The Tech Stack
Whatās striking about Thomasā build isnāt how technical it is ā itās how simple.
No custom platform. No heavy engineering. Just a thoughtful combination of tools, glued together by first-principles design:
ChatGPT / Custom GPTs ā OILSmith & FutureSmith agents, the brains of the system. Lightweight prompts, trained on Juroās frameworks, that turn feedback into practice and conversations into forward-looking plans.
Humaans ā the HRIS backbone, feeding employee data and workflows into the loop.
Zapier ā invisible automations ā logging outputs, sending reminders, closing loops without extra clicks.
Slack ā nudges and reminders delivered where work already happens.
Google Sheets ā the simplest possible reporting layer. A lightweight data stream, not another dashboard.
Notion ā the āOrg Brain,ā a living hub where frameworks, logs, and learnings all connect.
The takeaway? You donāt need a vibe-coded platform to change performance.
Sometimes, a few agentic GPTs stitched into everyday workflows can unlock massive value: saving 100+ hours, near-100% adoption, and creating cultural momentum that traditional review platforms never achieve.
š” The Wins
The numbers speak:
Adoption: Near 100% completion ā 300+ OIL conversations in 2 months.
Time saved: CRO estimates 100+ hours freed up.
Cultural shift: Faster decisions, fewer lingering issues, and the potential to eliminate PIPs altogether.
Proof: Internal Slack shout-outs + 4.8 GPT ratings show managers actually like using it.
š Reflection & Lessons
Donāt fear mistakes. AI is āplaygroundyā ā iteration beats perfection.
Titles donāt matter. Even Juroās CTO asked Thomas how he built it. Everyoneās learning AI together.
Critical thinking > tools. Not everything has to be a vibe-coded platform. Simple, agentic solutions can be just as powerful.
š Wrap & Whatās Next
Thomas proved whatās possible when HR leaders design systems instead of inheriting processes.
Bold takeaway: Performance management is being rewritten. With AI, you can eliminate PIPs, accelerate improvement, and future-proof roles.
š Subscribe to Field Notes so you donāt miss the next build.
š Join the AMA with Thomas next Wednesday, September 10.
š Forward this to a People Leader still stuck in six-month reviews.