AI Mastermind AfterShow · September 22, 2026
What happens when the tool is more powerful than the task? In this week’s AfterShow, Joe Moore and Katrena Drake brought Jeff Valin, Chad Cook, and Tyler O’Neal into one conversation about AI cost, speed, skill, and responsibility. Jeff’s running image gave the room its title: stop driving your Ferrari to the mailbox.
What you’ll discover
Why a small, repeatable task can become expensive when the model changes.
How to decide whether a task needs AI, ordinary software, or both.
What teams must learn before they scale AI use.
Why shared knowledge and data rules matter more than another shiny tool.
The central question: What is the job?
Katrena opened with a practical question for each guest: what does the Ferrari metaphor mean in their own work? Tyler, who builds with AI at Egmer Marketing, described using a lighter model for a simple task rather than paying for a coding or design model where it offers no advantage. Chad took the question back to the result: define the problem before choosing the tool.
“what is it you're trying to accomplish.”
Chad Cook · 9:18 in the September 22 AfterShow
That is the hinge. A recurring contact upload may need a short script rather than a model making the same decision every day. AI can help write and test that script, then step out of the way. Chad’s point was not that smaller is always better; it was that the system should fit the job, the volume, and the responsibility that comes with it.
The cost is in the workflow, not only the prompt
Jeff told a story about an automation designed to run a few times a day. Its original model choice kept each run inexpensive. Someone switched to a more costly model and then ran the workflow across a large batch of contacts. The bill rose quickly. Jeff said the batch also generated hundreds of leads, but the fortunate outcome did not erase the design lesson: understand the cost of a decision when it repeats at scale.
“understanding the tool, understanding the costs, is important.”
Jeff Valin · 3:42 in the September 22 AfterShow
Chad added that a prototype’s cost is different from a product’s cost. Multiply a task by users, transactions, and frequency before deciding what belongs in the production workflow. Tyler brought it back to a familiar engineering move: let AI help create a reusable tool when the task is stable and repetitive.
“There's no reason to use an AI that does that repetitively because you're wasting tokens”
Tyler O’Neal · 10:46 in the September 22 AfterShow
New tools do not remove old responsibilities
The room did not stop at model prices. A new product or agent may promise speed, but someone still owns its access, data, and maintenance. Tyler compared adopting AI tools to installing software plugins: the organization needs to understand what it has brought in and keep it current. Chad argued that teams need a shared answer to what information can enter a system, where it sits, and who is coordinating overlapping work.
Jeff pointed toward the skills and practices that make a team distinctive. General knowledge and code are increasingly available. The way a business understands a client, drafts a proposal, or makes a judgment is harder to copy. The useful question becomes how to share that know-how safely without losing the human decisions inside it.
Five signals from the room
Start with the outcome. Describe the task before choosing a model.
Count the repetitions. A cheap single run can become an expensive default.
Write the stable parts down. Reuse code and documented processes where they fit.
Teach the team. Accessibility makes building easier, not ownership optional.
Protect the special sauce. Capture the judgment behind the work without exposing what should stay private.
The Ferrari is real. It just does not need to make every trip. The people in the room are doing the harder work: knowing which trip matters.
Join the next AI Mastermind · Put your name in the directory · Join AI Posse for future replays
