The Missing Playbook for Agentic Systems
Why value doesn’t live in the model— it lives in the workflow.
Over the past year, I’ve been in countless conversations with teams trying to understand where AI agents actually create value. Not in demos, not in prototypes, not in impressive model outputs—but in the messy, pressure-tested reality of real operations.
Everyone feels the shift happening.
Nobody feels confident about where to place the pieces.
This audio is a walkthrough of the framework I use when working with product, design, data, and operations teams. It’s the most complete explanation I’ve given publicly so far about how agentic systems fit into real workflows, how to discover where they belong, and why product discovery—not prompting—is the real center of gravity.
If you’re wrestling with where agents fit inside your organization, I think this will help you.
Listen now!
Why I Recorded This
There’s an enormous gap between what AI tools can do and where they should actually participate inside an organization.
Most teams start at the wrong angle:
“What can the model do?”
But capability doesn’t tell you where value is.
Value appears where human judgment accumulates:
Where someone hesitates…
Where context is needed…
Where time quietly disappears…
Where the cost of being wrong is felt…
That’s where agentic systems belong.
In the audio, I break down how to identify these “moments of strain,” how to evaluate whether an agent should act, and how workflows need to adapt to make this shift work.
What’s Inside (The Short Version)
1. Why software is no longer just a surface
We’ve crossed a line: systems can now interpret, decide, act, and escalate.
This changes what software is allowed to do in the workflow.
2. The real person who matters
Every workflow has someone who carries the undocumented judgment.
Agentic systems don’t replace this person — they scale what they know.
3. Why demos are a trap
Most orgs build demos first.
Demos tell you nothing about where value lives.
4. How to find the exact moments agents belong
These are the “strain points” — cognitively heavy decisions or high-volume decisions.
5. How services turn into products
When judgment becomes legible and workflows break into agentic steps, entire categories of service work become scalable product value.
6. How experts train the system
Not with datasets — with corrections, explanations, and real cases.
Evals are just examples of correct reasoning.
7. How this changes organizations
Leaders need to think beyond demos and protect the space for workflow reinvention.
Why This Matters Now
We’re at a moment where organizations who understand agentic workflows will leap ahead — not gradually, but structurally.
The competitive advantage isn’t:
access to models
prompt engineering
or flashy features
It’s the ability to place judgment in the right parts of the workflow…
and scale it.
If you want a practical, grounded way to think about that shift, this audio is for you.
If This Resonates
Send it to someone on your team who keeps asking the right questions.
Or reply — I always enjoy digging into real workflows and helping teams figure out where agents actually fit.
Thanks for reading — and listening.
James