Range Podcast

Tobias Lipsky, founder and MD of TML Investments and Consulting, explains what IT is weighing when it says no to pay data in AI, and the questions to bring into that meeting.

What if AI could act as a team of challengers, not just a tool that completes tasks? Yvonne Prang joins Giac Soliman to explain how she built a “digital board” of specialized AI agents to analyze strategy, finance, M&A, and people questions. She shares how she tests their answers, checks sources, and uses human judgment to improve the work. The conversation also explores why leaders need AI fluency and how governance can help promising experiments scale.

A rewards leader needed budget scenarios for the board, and the pay review tool couldn't model them.

Nilan Vagh is Associate Director, People Operations and Systems at Zip, and instead of another mirage of spreadsheets he built the budget model himself in Google Apps Script.

He walks through it live on a fictional 250 person company, from a compa ratio heat map that feeds straight into the budget scenarios to the audit trail that makes the model defensible.

What we get into

Why ChatGPT and Codex were the accelerant and not the architect, with the business logic coming from his rewards lead. Why he builds almost everything in Google Apps Script, where a shared link won't open outside the company's Google login. Why he doesn't think vibe coding will replace HR software, at least for now, when people data needs security you can't vibe code. What he wants to build next, a rewards intelligence view that flags high performers whose equity has finished vesting.

About Nilan

Associate Director, People Operations and Systems at Zip, where his focus is increasingly shifting towards people infrastructure, governance and AI enabled workflow design. He builds internal products across People and Culture that turn fragmented, judgement heavy work into governed decision support. He speaks here in a personal capacity.

Four year equal vesting with refreshes on top looks generous until year five, when it drops off a cliff.

Colette Leung is Head of Total Rewards at Chainlink Labs, and she built a simulator to show what a new joiner actually experiences rather than what the grant letter says.

She models the alternatives live, including the front loaded schedules she thinks Robinhood and others have explored, against a market competitive line.

What we get into

The first thing she ever built with AI, which was a working manual of her manager. Why she thinks the development work has shrunk from about 70% of a build to maybe 10%. Moving long term incentive off Google Sheets into a system carrying 1,200 users. What she would fix with a magic wand, which is quantifying the difficult to quantify.

About Colette

Head of Total Rewards at Chainlink Labs, based in the Netherlands. She transformed the long term incentive programme and built an in house equity platform supporting 1,200 plus users. 15 plus years across high growth technology and consulting, including Uber and Deloitte.

Show notes at range.community.

Martin Smit spends his weekends building whatever has caught his attention. a gardening site, an art site with 200,000 pieces on it, etc.
The same habit has now read the annual compensation reports of the world's 10,000 largest companies, for 500 dollars a month.

Greg Laney has spent over twenty years between compensation and HR technology, most of it running Workday and HRIS strategy before pivoting back into comp roles. Job matching is the part of the work almost nobody says they enjoy. A manager sends over a paragraph, and somebody has to turn it into a survey match, an internal level, a salary grade and an explanation that manager will accept.

He built the thing that does it. It started as a ChatGPT agent at home that wasn't good enough, became a Copilot version at work that cut four hours a day down to about ten minutes, and is now an n8n workflow he put together in roughly four hours. Submit a job description and it checks the market survey, the internal job catalogue and the salary structure, returns a primary and a secondary match with a confidence flag, drafts the email to the manager, and writes the whole run back to a spreadsheet so anyone can see six months later why the job was graded the way it was.

01:38 Recruiting, Hay points, and the mentor who taught him Access

03:10 Where to start when work only gives you Copilot

10:20 Job matching, the job nobody says they love

13:49 Screen share, the workflow walkthrough

24:15 How n8n reaches your other systems

29:01 How does it know your pay philosophy

32:37 The judgement AI hasn't caught up to

Theresa Cortese runs total rewards at Nirvana Insurance, with twenty plus years across retail, semiconductors, ed tech and HR tech. Her business operations team owns the metrics and can't see pay data, so every quarter two spreadsheets got checked against each other by five senior people. Too expensive to keep doing by hand, not big enough for an enterprise platform to make sense.

Her CEO had told the company that experimenting with AI wasn't optional and that everyone should build something that week. So she did. The system she shipped runs Google sign in at the database query level, compares any two calculation runs, generates payout letters explaining how each number was reached, and passed Q1 actuals for every employee. She had never opened Terminal on her Mac.

The frame she uses is the useful part. She treats it as an entry level headcount she finally got approved. She teaches it, checks it, and coaches it forward when it gets something wrong. She is clear that it cost her time before it saved any, and that her own knowledge of all eleven variable comp plans is what makes any of it safe.

00:00 Who Theresa is
00:55 How far she trusts the output
01:19 The first thing she built
02:39 Fifteen years of merit spreadsheets
03:10 Where the ideas come from
04:25 No mentor, a team that experiments together
05:10 Shipping V1 to IT
06:35 Version thinking and imposter syndrome
09:00 Whether Claude Code is actually hard
11:03 They process, they do not think
13:04 Does AI save time or create more work
15:34 Why startups underinvest in total rewards
21:14 The story behind the tool
24:24 Screen share walkthrough
30:03 Advice to her earlier self
31:40 What comp teams look like in a year

Arif Ender on why compensation teams need a product mindset, cleaner data, and stronger human oversight to make AI actually work in the real world.

Ryan Buhrke and Giac explore how to shift from dependency to empowerment with AI, especially if you’re ready to stop waiting for IT approval and start building.