What to Automate, What to Keep Human

Here’s the dilemma every program manager’s facing in 2025:

AI can save me hours — but if I automate too much, do I lose the soul of my program?

Startup programs are operations-heavy by nature. Applications. Onboarding. Calendars. Reports. Events. Follow-ups.

There’s no shortage of moving parts.

And now, with AI tools everywhere, we’re tempted to automate everything.

But here’s the truth:

You shouldn’t.

Because the best programs aren’t just efficient — they’re alive.

They earn trust, create magic moments, and build lasting relationships. That doesn’t happen through prompts and bots alone.

Today, let’s break it down:

• What to automate

• What to keep human

• And how to find your balance

Why Program Ops Is Prime for AI — and Where It Goes Wrong

AI thrives in structure: repeatable tasks, clear data, consistent outputs.

Program ops is full of those.

But AI fails in nuance: emotional cues, cultural context, relationship building.

Which — surprise — is also a big part of running great programs.

So here’s the guiding principle:

Automate the repeatable. Humanize the relational.

What to Automate (And the Tools to Use)

Start with the bottlenecks.

These four areas are where AI gives you the biggest ROI:

✅ Application Management

• Scoring entries based on keyword relevance

• Auto-tagging founders by stage/sector

• Sending follow-up emails

Tools: Notion AI, Airtable + Automations, ChatGPT API

✅ Scheduling & Reminders

• Coordinating mentor sessions or office hours

• Sending calendar invites and reminders

• Rescheduling automatically when things shift

Tools: Calendly, Motion, Reclaim, Zapier

✅ Data Collection & Reporting

• Aggregating feedback from surveys

• Tracking engagement metrics

• Auto-generating visual dashboards

Tools: Google Forms + Looker Studio, Typeform, Notion databases

✅ Content Drafting & Summarizing

• Transcribing sessions

• Creating short LinkedIn recaps or email digests

• Drafting follow-up emails post-events

Tools: Otter.ai, Notion AI, ChatGPT (custom GPTs work well here)

Pro Tip: Let AI draft — but always review and edit before publishing. Your voice matters.

What to Keep Human (Non-Negotiables)

Some things don’t belong in the hands of an algorithm.

❌ Startup Onboarding

This is the emotional handshake of your program. Founders need to feel seen. Not just processed.

❌ Mentorship Matching

Let AI suggest. But final pairings need human judgment. Chemistry matters.

❌ Conflict Resolution or Program Coaching

These moments require empathy, nuance, and trust. No prompt can replicate that.

❌ Cultural Calibration

Running a program across different markets? Language, behavior, and unspoken norms vary wildly. Don’t risk misunderstanding by automating this.

If it builds trust or sparks creativity, keep it human.

Finding Your Balance — A Simple Framework

Don’t aim for “AI-first.”

Aim for hybrid excellence — the best of tech and touch.

Here’s a simple filter to apply before automating anything:

The 3C Test

Ask: Is this task...

• Clerical? → Automate it.

• Creative? → Co-create with AI, but add your voice.

• Conversational? → Keep it human.

And every quarter, audit your workflows:

• What’s draining your time that AI could handle?

• What’s lost its human spark?

Make small changes. Test. Iterate.

The Takeaway

AI won’t replace great program managers.

But great program managers who use AI wisely will outperform the rest.

This week, do this:

✅ Identify one task you're doing manually that AI could automate.

✅ Identify one moment in your program that needs more human connection.

✅ Make one small change to improve both.

Efficiency is the baseline.

Connection is the differentiator.

Talk soon,

Yaniv