Reengineering HR with AI
Week 3
A high performer spent eighteen months in client management, doing good work but clearly not thriving. What we didn’t know at the time, and what she didn’t know how to articulate, was that she wanted to do strategy work. Because we didn’t know, when a role opened up in business strategy, we filled it with an external candidate.
By the time we figured out she was in the wrong seat, we’d wasted a year and a half of potential, and she’d already started looking elsewhere.
Most (not all) managers have no idea what their people are actually capable of. Not because they’re bad managers, but because they’re drowning in the day-to-day work of keeping projects moving and putting out fires. Strategic people development, the kind where you really understand someone’s knowledge, skills, and abilities, falls to the bottom of the list.
I’ve seen this pattern play out dozens of times.
Someone’s great at their job, so we leave them there. We don’t ask if they could be great at something else. We don’t check if they’re building new capabilities. We don’t track if their strengths have shifted. We wait for the annual performance review, fill out the form, and move on.
Meanwhile, HR tells leadership we need better succession planning and talent mobility. We create competency frameworks and skills matrices. We talk about playing to people’s strengths. But we’re making all these decisions based on data that’s six months old, or twelve months old, or based on the role someone was hired for three years ago.
The only time we get a real update on someone’s KSAs is during the annual review cycle, when managers are rushing to complete fifty forms in two weeks. That’s not assessment. That’s checkbox compliance.
Here’s what happens when you only assess people once a year.
We had a critical project that needed someone with strong analytical skills. Leadership wanted to move fast, so they brought in a consultant. Expensive, but they delivered.
Four months later, I was reviewing development plans from the annual review cycle. Buried in the comments, a manager had noted that one of his direct reports had been asking for more analytics work and had completed a business analytics certification on her own time. The timing lined up perfectly with when we’d hired the consultant.
I called the manager. “Why didn’t you tell us she was developing these skills?” He was genuinely confused by the question. “I noted it in her review. I figured if HR needed to know, they’d ask.”
We’d paid a consultant $80,000 to do work we had someone ready and eager to do. She didn’t even know the project existed. By the time we connected the dots, she’d accepted an offer at another company where they’d noticed her potential faster than we did.
I understand why this happens. Managers are busy.
Really busy.
Asking them to provide quarterly updates on their team’s KSAs feels like one more thing on an already impossible list. And let’s be honest, most managers don’t know how to assess skills in a way that’s useful to HR. They know if someone’s doing their job well, but translating that into “what else could this person do?” requires time and thought they don’t have.
There’s also the fairness concern. If we’re assessing some people quarterly, shouldn’t we assess everyone? And if we’re collecting all this data, what’s our obligation to act on it? What if we identify someone’s strengths but don’t have a role for them?
These are real constraints. But they’re not reasons to stick with a system where we only formally check in on people’s development once a year, when it’s already too late to make timely decisions.
Taking this problem, I had an idea to reengineer the solution. What if we used an AI agent to gather quarterly KSA assessments from managers.
Set up an agent that reaches out to managers every quarter with a simple prompt. “Let’s talk about your team’s development.”
It asks preset questions, like “Who’s developed new capabilities in the last three months?”
“Who’s ready for more complex work?”
“Who’s showing strengths outside their current role?”
“Who’s struggling and needs support?”
The manager responds conversationally. The agent follows up with, “You mentioned Sarah’s showing strategic thinking. Can you give me a specific example?” The manager recalls, “She restructured our client onboarding process and reduced time-to-value by 30%. She didn’t just execute, she redesigned the whole approach.”
That response gets logged. Not in a performance review form that sits in a system until next year, but in a spreadsheet that HR reviews monthly. Now when a strategy role opens up, you’re not guessing who might be ready. You have recent data showing Sarah’s already operating at that level.
This isn’t surveillance. Managers aren’t being asked to do more work, they’re having a five-minute conversation with an agent instead of trying to remember twelve months of performance in December. And it’s not replacing the annual review, it’s making that review actually useful because you’re working with current information, not stale observations.
Start with one team next month. Build a simple agent. Write a prompt with five questions about team development. Have the agent ask the questions, capture the responses, and put them in a spreadsheet. Review it with leadership and ask: “What decisions could we make differently if we had this data every quarter?”
I’m betting you’ll identify at least two people who are ready for different roles and one critical skill gap you didn’t know existed.
If you’re only formally assessing your people once a year, how many opportunities are you missing while you wait?
How many high performers are sitting in the wrong roles because no one’s checking if their strengths have evolved? How many external hires are you making when you already have someone internal who’s ready? How many people are leaving because they developed capabilities you never noticed?
You can’t claim to care about talent development if you’re only paying attention one month out of twelve. You can’t build a high-performing organization on year-old data.
AI agents aren’t adding bureaucracy. They’re creating the visibility you’ve always claimed you wanted but never had the capacity to build. The question is whether you’re going to use them or keep pretending the annual review cycle is enough.
What are you actually going to do differently?