The Track
A Section Blog
The hidden reasons you’re not getting AI ROI
AI is good enough, the humans need help
Newsflash if you’re waiting for AI to get better before you invest: It’s already pretty good, it’s the people using it that need to get a lot better – and fast. Greg is taking over this week’s newsletter to show you why.
When to use a specialized AI tool vs. an LLM
If you’re hitting the limits of your LLM because you need more data or niche training, you’re likely doing a lot of high-end knowledge work – and that’s where specialized AI tools come into play.
We tested two Deep Research tools. One was unusable.
Two LLMs have released Deep Research features without much of a splash. So if you’ve been wondering about them (and don’t want to pay the $200/month price tag), read on for our Chief of Staff’s take on ChatGPT vs. Gemini.
How an agency is preparing for AI clients
AI has put a big question mark over the future of agencies. So we talked to David Freas, Managing Partner at Supermoon, about how he’s transforming his brand agency with AI agent teammates – and for AI agent clients.
What you really need to know about ChatGPT-5
ICYMI, OpenAI launched ChatGPT-5 . If your head is still spinning from the other 6 models they’ve released this year, here’s our guide on what’s different and what you actually need to pay attention to.
You're bad at using AI because you're bad at managing people
Managing people and managing AI are the same skill. So if you can’t get a good output from AI, slow down and treat it like a direct report.
Squishy vs. hard ROI: Why leaders need both
One of the hardest parts of AI ROI is figuring out how to report qualitative wins to execs who want quantitative reports. At our AI:ROI Conference, Michael Domanic shared the framework for approaching these two disparate kinds of ROI.
AI’s progress may be holding your team back
On paper, the rapid release of new AI models and features looks like a win for knowledge workers. In reality, teams are drowning in AI overwhelm. Here’s how leaders can help them scale an ever-steeper learning curve.
