GLN 2024 - AI Driven Needs Analysis

[Paul Terlemezian]

Wednesday, December 11, 2024 at 3:00 PM

Session Leader: Bill Crose

https://us02web.zoom.us/j/89895585667?pwd=nSM3X2YA87Cj5CbPl0N3SCZvR...

• Meeting ID: 898 9558 5667
• Passcode: 086728

The big picture is that innovation, as determined by annual USPTO grants, is accelerating at around 11% per year. We saw 3M patents in the 40 years from 1979-2019 and will see that number balloon to 88M in the coming 40. Each of those patents represents at least 1 product development job and likely exponentially more. Few of those jobs will be permanent and most workers will migrate from innovation to innovation.

Training and learning will be ineffective at enabling future workers to do their jobs because there will be too many SOPs changing too frequently for anyone to memorize. Pythia is how businesses can enable, assure, and continuously improve performance. 

One of the big problems/opportunities that AI can resolve is needs analysis. Every training initiative is supposed to begin with a proper needs analysis, but I'm guessing nobody in the group has ever performed a "proper" one. We're lucky if our clients all allow us to do any analysis at all. Needs analysis is how managers, directors, and executives spend their time. Managers presumably spend their time reviewing data, applying their expertise/sophisticated guessing, then choosing what problems to address via which interventions. AI can mimic how managers/directors/execs perform needs analysis and reveal needs faster, better, and cheaper than humans because it doesn't insert bias, greed, or self-preservation into its decision-making process.

From what I've found so far, all it would take to make AI perform company-wide needs analysis, not just training/learning, is a workflow, algorithms, and data (valid/unbiased/reliable/performance ratio data). I pasted the workflow into the presentation but it's all pixelated, so here's a link to it: docs.adytonusa.com/ANAM.jpg

What I'm thinking for the presentation is to pick up where we left off this week with what AI and IA are, what I think will happen to jobs when AI matures, and how AI will impact learning and performance. I'd make the case for automating needs analysis by pointing out some of the bias that impacts human-performed analysis, why workers should prefer AI to analyze needs, and that businesses could reap a windfall by replacing highly paid managers/directors/execs with AI-driven needs analysis. From there, I'm thinking of talking through the workflow and stimulating conversation about the next steps after finishing the workflow, i.e.: what data is needed, what databases, how to capture the data, what algorithms are needed, what the algorithms would do, etc. At the end, I'm naturally hoping to get good feedback, but am mostly hoping they'll agree that enabling AI to perform needs analysis would be a good thing and that they can envision how simple it would be to write the algorithms, develop the datasets, and implement.