It's impossible to escape the noise surrounding Artificial Intelligence.

From breathless headlines about job-ending automation to confusing debates about chatbot sentience, the hype is overwhelming. In this environment, it's challenging to know what's real, what's speculation, and what truly matters for your work.

But here's what I've learned from conversations with world-class experts on the "Directionally Correct" podcast: most companies are approaching AI precisely the wrong way.

They're starting with job analysis. They're asking, "How will AI change this role?" They're chasing five-minute productivity wins like meeting summaries.

And they're missing the entire point.

The Problem With Starting Small

Cole Napper and Scott Hines recently marked their 150th episode. No single topic has dominated their conversations more than AI's impact on the workplace.

Cole's take is direct: if he were leading a people analytics function today, he'd put all his effort into making the company more innovative. AI is the essential tool to achieve that goal.

Scott describes the innovation process as an "oscillation" between exploring new ideas externally and refining them internally. This is precisely the cycle that AI is set to turbocharge.

Not efficiency. Innovation.

What The Experts Are Actually Doing

Prasad Setty was VP at Google and founded its people analytics team. He's not using AI to summarise emails.

He's built a virtual brain trust.

Missing the daily collaboration of his former team, Prasad recreated the personas of his key team members within ChatGPT and Gemini. When he brainstorms, he doesn't just ask the AI for its opinion. He asks his virtual team.

Each persona offers a unique perspective—one focused on execution, another on people. It creates a dynamic, multi-faceted dialogue.

"We could all play that in our own minds, but it becomes too cognitively difficult. At some point, you lose who you are and try to see the world from very different perspectives. Gen AI can play a helping hand and help you build your own perspectives."

This isn't about better brainstorming. It's a practical method for democratising strategic thinking and mitigating the groupthink that plagues even the most talented teams.

Prasad draws a sharp distinction between convergent problems (finding a single right answer) and divergent problems (exploring many possible ideas). He sees AI less reliable for convergent tasks due to hallucinations. But for divergent challenges? It's "handy" for broadening your initial perspective and generating a wide range of possibilities.

When the 277-page judgment in the Google antitrust lawsuit was released, Prasad didn't wait for media summaries. He used Gemini to read and analyse the document himself. AI empowers you to become your own intelligence analyst. You can cut through media bias to understand primary source information for yourself.

The Army Gets It Right

Kris Saling is the Director of Talent Innovation for the U.S. Army. She's implementing AI in a massive organisation where profit isn't the goal.

So how does she measure ROI? Not in dollars. In readiness.

Her team quantifies AI's value by calculating "the amount of minutes we can put on mission." By automating administrative tasks, they give soldiers back valuable time to focus on their core competencies.

This reframes the entire conversation. It moves the goalpost from cost savings to mission amplification.

For any organisation—non-profit, educational, or corporate—this model forces you to ask a more important question: not "how much money did we save?" but "how much more of our core purpose did we achieve?"

The Army frames its AI strategy around change management. The focus is on "evolving the tasks that are being done versus replacing the jobs."

As low-value tasks are automated away, individuals are freed to take on new, more critical responsibilities. Automation becomes a logical evolution of your role, not a threat.

Kris argues that strategic workforce planning "has to include the machine." AI should be viewed not merely as a tool that augments human processes, but as a fundamental component of the workforce itself. When planning for the future and asking "who does the work," the answer will increasingly involve a partnership between humans and machines.

The Right Sequence

Alexis Fink is a veteran people analytics and workforce strategy leader from Meta, Microsoft, and Intel. She provided the most precise blueprint for how organisations should approach AI.

AI should free analytics teams from the "tyranny of dashboards." Stop building bespoke reports for every executive whim. That's low-value work.

Companies are approaching AI integration in the wrong order. They shouldn't start by analysing how AI will change a specific job. Here's the correct sequence:

1. Rework the entire business process to be more efficient and effective
2. Analyse the tasks within that new process to identify what can be automated
3. Determine whether a person, a machine, or a human-machine partnership is best suited for each task
4. Then, and only then, perform a job analysis to build the new job description

The job analysis comes at the end, not the beginning.

This process-first, people-second approach ensures that technology serves a reimagined workflow rather than simply automating the inefficiencies of an old one.

Why AI Pilots Fail

Many corporate AI pilots are failing because they're focused on the wrong things.

Companies are chasing minor personal productivity gains—summarising meeting notes, drafting emails—instead of redesigning core business processes.

Alexis points to the "naive assumption" that individual efficiency automatically translates into organisational efficiency. A collection of individuals saving five minutes on a task doesn't necessarily make the business more effective or profitable.

True transformation requires rethinking the work itself.

Both Kris Saling and Alexis Fink push us to redefine how we measure AI's value. Saling's focus on "readiness" and "minutes on mission" offers a strategic alternative to mere efficiency gains.

The goal isn't just to save individual time. It's to amplify organisational impact and achieve more of your core purpose.

The Real Opportunity

The true potential of AI lies not in automating old tasks, but in reimagining our work entirely.

It's a shift from incremental efficiency to fundamental effectiveness.

Alexis Fink provides the macro-level vision: redesign core business processes from the ground up. Prasad Setty offers a powerful micro-level tool to execute that vision: use AI to build a virtual brain trust that can challenge assumptions and generate the divergent ideas needed for true transformation.

The leaders who grasp this will move beyond the hype and into a future of genuine innovation.

Instead of asking how AI can summarise your notes, ask how it can help you build a virtual brain trust to redesign a core business process completely.

Want more insights like these? Subscribe to the Directionally Correct podcast to keep up with interesting perspectives and candid coffee talks with experts like Prasad Setty, Kris Saling, and Alexis Fink. Hosts Cole Napper and Scott Hines cut through the noise to bring you practical strategies from the front lines of technology, government, and business. New episodes drop weekly—subscribe wherever you get your podcasts.