What I got wrong about AI in HR - and what finally worked
Summary: Souleimen Ach, an HR professional with a tech-savvy background, shared insights on their journey of implementing AI agents in the People function. Initially, they faced challenges by choosing unsuitable platforms and improper framing of AI benefits, which led to a lack of adoption. They learned to focus on integrating AI into existing workflows, measuring efficiency in terms of removed work, and leveraging early adopters to promote the solution. As a result, they successfully implemented AI agents for various HR tasks, enhancing overall strategic productivity without requiring more time. They invite others to explore their solutions and participate in a training session.
I'm a HR professional and a tech-savy person and when I started building AI agents for our People function, I made every mistake:
Picked a flashy platform that forced my team to leave Gmail. Adoption: 0%.
Framed it as "save HR 20% time" instead of "remove entire categories of work."
Treated chatbots and real agents as the same category. They're not.
Tried to convince skeptics with demos instead of letting one early win do the talking.
What finally worked:
Pick the ONE workflow eating your team's week
Build the agent inside the tool they already use — never a new platform
Set the bar at "good enough to ship with 30 min review," not "perfect"
Measure removed work, not added work
Let early users sell the rest of the team. Don't pitch internally.
We now have agents handling onboarding plans, JDs, performance scoring, and L&D kit production. Senior team's strategic output is up, not because they have more hours, but because their hours aren't fragmented anymore.
Want to go deeper?
🔗 See what we're building → aiydou.com
🎁 Test all 4 agents free (first 20 sign-ups, no credit card) → aiydou.gumroad.com/subscribe
📅 Live training session — "Deploying AI agents in HR, end to end" → https://www.linkedin.com/events/7459682079761543168/
What's the ONE workflow you'd kill first if you could?
