Great Mondays Radio, episode 173, with Josh Levine. Published October 9, 2026 on Substack. Josh’s show notes, reproduced verbatim.

Tim Shea is founder and CEO of LatticeWork Insights, a data science agency for retail and direct to consumer brands. Everyone says data or compute is the new oil. Tim thinks it’s domain expertise, because when your best, average and weakest people use the same AI, they get wildly different results.

My biggest takeaway: same tool, same license, wildly different results. AI shows you how far apart your people’s expertise already was.

My core reframe: most AI rollouts assume the tool raises everyone’s capability, so the job is getting people to use it. Tim sees the opposite. The output depends on what the person brings. His above average people become 10x engineers and 10x marketers. His average people spray AI onto every deck and email, and he has colleagues who plugged their Slack into Claude so it answers for them, with their face on it. The below average people fill the company with slop. Ask them a question and they paste it into Claude and send back whatever comes out. Tim thinks that’s a big piece of why companies aren’t seeing a return. Without expertise, you can’t tell when the AI is smarter than you and when it’s dumber, so you end up hitting reload and asking it to do better (his favorite example is the coffee that ended up with the Milky Way around it). Even the people who know what they’re doing pay a price. One engineering team he knows is writing 10x more code and is exhausted from checking 10x more edge cases. That’s the uncomfortable part for leaders. Buying everyone a Claude license spreads the tool. It doesn’t spread the expertise. If your AI plan doesn’t build and reward people who know the work, you’re paying to scale the slop.

About Tim Shea

Tim Shea is founder and CEO of LatticeWork Insights, a data science agency that helps retail and direct to consumer brands reach their next stage of growth. He’s run it for about a decade on top of 25 years in the field, with clients like the NBA, the UFC, Sweetgreen, Princess Cruises and Reddit. AI changed how he sells it. He now pitches return on analytics spend, because he says clients should judge him on whether he made them more money, not on whether they got a Tableau dashboard.

Listen and read the transcript on Substack →