Todd McLees recently shared something that caught my attention: "AI Agility is coming home to Milwaukee." He added that after two years building AI capability programs with higher education partners nationwide, this launch feels "a little more personal" because Milwaukee is home.

That mix of urgency and grounded pride is worth sitting with. In a world where AI announcements often feel abstract, Todd is pointing to something concrete: capability building that helps real people do better work with AI, at scale, inside real organizations.

In his post, Todd also drops a line I keep coming back to: "the gap isn’t technical." Most workplaces already have access to powerful tools. The constraint is what people can actually do with them, day after day, in the messy reality of business.

AI capability is not a tool problem

If you listen to most AI conversations at work, they quickly become tool conversations:

Which model is best?
Which vendor should we standardize on?
Which features can we turn on?

Those are valid questions, but they can become a convenient distraction. Todd’s point is sharper: access is widespread, but effective collaboration is not. People may have ChatGPT, Copilot, or a dozen niche apps, yet still struggle to turn AI output into trustworthy decisions, better workflows, or measurable value.

Key idea: When intelligence is abundant, discernment becomes the scarce resource.

That is why programs like AI Agility resonate. They are not primarily about teaching prompts or showcasing features. They focus on the human skills that make AI useful and safe in the first place.

What "AI Agility" is really aiming for

Todd describes AI Agility as a Human x AI collaboration program built with input from thousands of professionals across sectors and dozens of institutional partners. That breadth matters because the skills that transfer are not industry-specific. Finance, healthcare, manufacturing, education, and professional services all face the same core challenge:

AI can produce options quickly.
Humans must choose, refine, verify, and act.

From a practical standpoint, AI Agility is about building repeatable capability. Not a one-time training, not a motivational keynote, and not a pile of AI policies nobody reads.

Todd highlights several themes that should be front and center in any serious AI upskilling effort:

Judgment matters

AI is excellent at generating drafts, summaries, patterns, and possibilities. But judgment is what turns possibilities into decisions. Judgment is also what catches subtle errors, bias, context mismatch, and false certainty.

In many roles, the cost of being wrong is higher than the cost of being slow. That is why the best AI-enabled teams are often the ones with the strongest review habits, not the flashiest automation.

Augmentation often beats automation

Automation is appealing because it promises headcount-free productivity. But it can fail quietly. An automated step that is wrong 5 percent of the time can create downstream chaos, rework, and mistrust.

Augmentation is usually the healthier starting point:

Use AI to prepare a first draft.
Use AI to generate alternatives.
Use AI to find gaps and edge cases.
Use AI to help you explain your reasoning to others.

That approach keeps humans accountable while still capturing speed and breadth.

Direct intelligence to create value

Todd’s phrasing here is subtle and important. Having intelligence available is not the same as applying it well. Value is created when teams can aim that intelligence at the right problems, with the right constraints, and the right definition of success.

This is where strategy meets execution. If a team cannot articulate what "good" looks like, AI will not fix that. It will only produce more output.

Why the "skills gap" is socio-cognitive

Todd references research from Northeastern and Harvard that supports a powerful claim: the skills that drive successful Human x AI collaboration are the same socio-cognitive skills that drive successful collaboration between humans.

That rings true. Think about the best cross-functional collaborators you know. They tend to:

Ask clarifying questions
Surface assumptions
Seek feedback early
Communicate tradeoffs
Separate ideas from identity
Update their view when new evidence appears

Now map those to AI work:

Clarifying questions become better problem framing.
Surfacing assumptions becomes better constraints and evaluation criteria.
Seeking feedback becomes iteration with structured review.
Communicating tradeoffs becomes transparent decision-making about using AI output.

In other words, the future of AI work is not just technical fluency. It is collaborative fluency.

If your team struggles to collaborate with each other, they will struggle to collaborate with AI.

The Milwaukee launch: a practical pathway, not a hype cycle

Todd outlines a clear sequence for Wisconsin employers:

A free event on February 24 at Marquette: an "honest conversation" about what it takes to work effectively with AI at scale.
The first Marquette cohort launching March 9: the AI Agility Challenge.

He also gives specifics that signal this is meant to be operational, not theoretical:

20 modules
cohort-based and self-paced structure
a hands-on workshop
a certificate from Marquette Executive Education
a full year of community access with updated modules and events

Those design choices matter because capability building is a behavior change problem. People do not become effective with AI because they watched a webinar. They improve when they practice, compare notes with peers, get feedback, and keep learning as the tools and norms evolve.

Todd notes this will be their 15th release in 16 months, shaped by feedback from thousands of participants. That pace is a feature, not a flex. In AI, the content must evolve, but the underlying human skills can be reinforced and deepened.

Not a chatbot: why a learning companion is different

One of the most interesting parts of Todd’s update is the inclusion of an AI agent "designed specifically to guide people through building the skills and human agency" required for collaboration. He is careful to say: not a chatbot, but a learning companion aligned with how adults develop capabilities.

That distinction matters because many organizations are stuck in a shallow loop:

Deploy a tool
Tell people to use it
Measure usage
Call it transformation

A learning companion suggests something more intentional: guidance, structure, reflection, and nudges toward better habits. In practice, that could look like:

prompting you to define success criteria before generating outputs
asking you to rate confidence and identify verification steps
encouraging you to capture reusable workflow patterns
helping you document what changed in your decision after reviewing AI suggestions

Those are the moves that build durable skill, not just temporary convenience.

What employers should take from Todd McLees’s message

If I had to translate Todd’s Milwaukee announcement into a playbook for leaders, it would be this:

1) Stop treating AI adoption as a software rollout

Treat it as capability development. That means structured learning, practice, coaching, and time.

2) Define where humans must stay in the loop

Decide where judgment is essential, where risk is unacceptable, and where AI can safely accelerate work.

3) Build habits that make collaboration sustainable

Sustainable Human x AI work requires repeatable behaviors: verification, versioning, documenting assumptions, and sharing patterns.

4) Invest in agency, not dependency

The goal is not that the tool feels magical. The goal is that your people feel more capable, more accountable, and more impactful.

That is the opportunity Todd points to when he says intelligence is abundant. The winners will not be the organizations with the most AI output. They will be the ones with the strongest human capability to steer it.

This blog post expands on a viral LinkedIn post by Todd McLees. View the original LinkedIn post →