1. Not all AI use cases are created equal

Pilar broke AI use into two buckets: speeding up tasks you already do, and supporting complex decisions. The first is straightforward. The second is where the real risk — and opportunity — lives.

“Automation is an equation: inputs, outputs, workflow. But when you're using AI to influence performance decisions or diagnose issues, the stakes are higher. You need to understand what the model can and can't do.”

She emphasized that HR leaders should use AI to assist, not replace, human judgment. For example: use it to find patterns in engagement data, but don't let it single-handedly assess who's underperforming.

2. AI can radically improve performance reviews — if you feed it the right data

Pilar shared how she and Ian (ChartHop’s CEO) are testing AI-generated reviews based on their ongoing documentation in ChartHop. Weekly notes, one-on-one recaps, project outcomes — all of it gets fed in.

“It wrote a self-review that was so accurate, it felt more honest than what I would’ve written myself.”

The key? Consistent data capture over time. If your managers and employees aren’t documenting regularly, the AI can only reflect what it sees — which may not be much.

3. Engagement survey analysis shouldn’t take 50 hours

At her last company, it took weeks to process engagement survey data. At ChartHop? Three hours.

“I analyzed all our qualitative and quantitative data, created an action plan, and even wrote a follow-up survey. That used to take 50-70 hours.”

The difference is generative AI. Instead of sorting through spreadsheets manually, you can ask the system to synthesize themes, identify hotspots, and even suggest next steps.

4. Think of your AI tools as your own HRBP

One of Pilar’s best mental models: treat your AI tool like the HRBP you’ve always wanted.

“Ask it the big, strategic questions. What should we automate? Who should be using AI? What efficiency gains are realistic? Then work with it like a partner.”

The more you feed it structured data, the more it can help you make decisions, run diagnostics, and operate like a force multiplier.

5. Don’t confuse friendly tone with good judgment

Just because an LLM sounds confident, doesn’t mean it’s right. Pilar warns against letting AI make nuanced people decisions without human oversight.

“It can be overly agreeable. One minute it's flagging someone as underperforming. The next, it says they’re great. You have to know its limits.”

Use AI to save time, explore patterns, and draft content. But for performance calls, org decisions, or sensitive topics? Your judgment still matters most.

See you next week!

P.S. If you like MPL, help us grow the show by giving us a 5 star rating on Apple or Spotify.