Let's rewind a bit. Back in my college days, I did the absolute bare
minimum to get my engineering degree. I wasn't the brightest bulb in
the class, and I genuinely wasn't interested in playing the whole traditional
engineering game.
But I had my favorites: Data Mining and the Theory of
Computation. I never really knew why I liked them at the time. But
looking back now, I think it was the abstractness—the thrill of finally seeing
patterns hidden in the noise.
I am constantly fascinated by technology and easily
distracted by trivia. Did you know that rather
than an engineering or mathematics journal, Alan Turing chose to publish his
foundational artificial intelligence paper in Mind, a leading peer-reviewed
journal focused on philosophy and psychology. A fact that I read in college
when I was reading about the Turing machine and since then it is living in my
mind rent free. Okay distracted again. 😅Coming back to the
learning nugget.
The L&D Brick Wall
Fast forward a few years into my career in Learning and
Development (L&D). I kept hitting the exact same wall.
Whenever the conversation turned to measuring the actual
behavioral impact of our programs, I was constantly told that measuring
behavioral skills is next to impossible. The accepted industry route? Send
out a 30-60-90-day post-assessment and hope for the best. It was all subjective
surveys and guesswork.
The Eureka Moment
Then the AI boom happened. As I did a deep dive into
its application in L&D, the "engineer" inside me suddenly woke
up.
It hit me: AI can create data, and data never
lies. This meant that all the training and the efforts can finally
be measured objectively. We don't have to rely on manager observations or smile
sheets anymore. We can measure behavior directly.
From Paper to Prototype
That Eureka moment led me down a rabbit hole to learn how to
build an agent.
The vision is to create a behavioral agent—my tool of choice
being Microsoft Copilot Studio. Instead of giving a trainee a multiple-choice
quiz, the architecture looks like this:
- The
Persona: The agent acts as a specific persona.
- The
Case Study: It runs the user through dynamic, conversational case
studies.
- Rubric
& Metrics: It uses built-in rubrics and metrics to measure
the trainee's responses and approximate the learning in the classroom.
We are no longer just hoping someone communicates better
after a workshop. We are putting them in a simulator and looking at the hard
data.
I do not know if it will be perfect right out of the gate,
and it is very much work in progress. But for the first time in a long time,
the abstract patterns are starting to make sense.
Comments
Post a Comment