There is a version of AI upskilling that a lot of organizations are doing right now: assign an online course, track completion rates, mark it done.
It feels like progress. It checks a box. It produces a number you can put in a report.
It is not transformation.
Catie Maillard, Global Head of People at Data Society Group, draws a sharp line between training and enablement:
“We’re not here to just train people so that you could check the box and say they know this tool. We’re here to help transform your entire workplace and help make sure that they’re implementing these skills and finding real business value in the tools”
WHAT IS AI UPSKILLING, REALLY?
At its most basic, AI upskilling refers to building the knowledge, skills, and confidence employees need to use AI tools effectively in their day-to-day work. It includes technical literacy and human competencies that make AI use genuinely valuable: critical thinking, judgment, workflow redesign, and the ability to decide when AI helps and when it gets in the way.
True AI upskilling is less about reaching a skill level and more about building the organizational muscle to keep learning continuously. For a deeper look at why most upskilling programs stall at the awareness stage, see AI Is Forcing Employees to Learn in Public (Whether They’re Ready or Not):
https://datasociety.com/ai-is-forcing-employees-to-learn-in-public-whether-theyre-ready-or-not/
THE TWO MODES OF EFFECTIVE AI TRAINING
The first is micro-learning for immediate, tactical application. Short, specific, searchable modules that answer: how do I do this thing with this tool right now?
“Folks either need right-now, point-of-time micro learnings. Like, I need a five-minute snippet around how I get Claude to do this.”
The second mode is instructor-led training for actual workflow transformation. This is where behavior changes.
“In terms of actual transformation, that is where the people touch is so important. Because when it comes to instructor-led trainings, that’s not just learning, that’s also enablement.”
For a thorough look at why instructor-led training consistently outperforms self-paced learning for AI adoption, see Why Instructor-Led Training Still Works (And Where Self-Paced Falls Short):
https://datasociety.com/live-instructor-led-training-vs-self-paced-learning-why-instructor-led-training-is-the-better-option/
WHAT ENABLEMENT LOOKS LIKE IN PRACTICE
Training delivers information. Enablement changes behavior. An instructor can diagnose where an employee is actually getting stuck, troubleshoot in real time, and help redesign a specific workflow.
“Someone can come in personally and help diagnose and help figure out and troubleshoot in ways that you can’t do yourself, or that you can’t do with Claude.”
WHY GENERIC AI UPSKILLING PROGRAMS FAIL
Generic AI training fails for the same reason generic anything fails: it does not meet people where they are. Effective AI upskilling has to be contextual, role-specific, and account for the emotional reality of AI adoption.
Organizations also need to rethink what their learning infrastructure can handle. Inc.: Why Organizations Must Evolve Their Legacy Learning Management Systems makes the case for why traditional LMS approaches are inadequate for the pace of AI change:
https://datasociety.com/inc-why-organizations-must-evolve-their-legacy-learning-management-systems/
“We need to give people the ability to experiment, and we need to give them the ability to experiment without repercussion.”
WHAT GOOD AI UPSKILLING PROGRAMS LOOK LIKE
The most effective programs are role-specific, combine on-demand resources with instructor-led workflow sessions, and measure success by behavior change and productivity outcomes rather than completion rates. For context on how upskilling and implementation must advance together, see AI Upskilling vs. AI Implementation: Why Enterprises Need Both: https://datasociety.com/ai-upskilling-vs-ai-implementation/
THE BOTTOM LINE
“The idea is to create practical courses that solve problems. We’re here for you to help transform your entire workplace.”
READY TO BUILD AN AI UPSKILLING PROGRAM THAT ACTUALLY WORKS?
Real AI upskilling changes how your people work, not just what they know. Data Society designs practical, instructor-led AI training programs built around your team’s actual workflows and tools. Reach out to find out what transformation looks like for your organization: https://datasociety.com/contact/
RELATED RESOURCES FROM DATA SOCIETY
– Why Instructor-Led Training Still Works (And Where Self-Paced Falls Short): https://datasociety.com/live-instructor-led-training-vs-self-paced-learning-why-instructor-led-training-is-the-better-option/
– AI Is Forcing Employees to Learn in Public (Whether They’re Ready or Not): https://datasociety.com/ai-is-forcing-employees-to-learn-in-public-whether-theyre-ready-or-not/
– Inc.: Why Organizations Must Evolve Their Legacy Learning Management Systems: https://datasociety.com/inc-why-organizations-must-evolve-their-legacy-learning-management-systems/
– AI Upskilling vs. AI Implementation: Why Enterprises Need Both: https://datasociety.com/ai-upskilling-vs-ai-implementation/
Frequently Asked Questions
AI upskilling is the process of building the knowledge, skills, and practical capabilities employees need to work effectively with AI tools. It includes technical competencies and human competencies like critical thinking, AI oversight, and workflow redesign. Genuine AI upskilling results in changed behavior and improved work outcomes, not just awareness of how AI tools function.
AI upskilling builds on existing skills to improve performance in a current role. AI reskilling prepares someone for an entirely new role or function. Both are important, but most organizations prioritize upskilling in the near term because it delivers faster productivity gains.
AI upskilling most commonly fails because organizations treat it as a training rollout rather than a change management effort. Generic, self-paced courses not tied to specific tools or workflows produce knowledge without behavior change.
Effective upskilling is ongoing rather than time-bound. A foundational program might take a few days of instructor-led training combined with ongoing micro-learning. Meaningful workflow transformation typically happens over weeks to months as employees practice, iterate, and receive feedback.
Look for providers who ask about your specific use cases, roles, and governance environment before designing a program. Be cautious of any provider offering a one-size-fits-all curriculum. Effective enterprise AI training is designed for your organization’s context, not adapted from a general consumer course.
An effective program includes foundational AI literacy, tactical micro-learning modules, instructor-led sessions tailored to each team’s roles and challenges, structured opportunities to practice in a psychologically safe environment, and feedback mechanisms that measure behavior change rather than just completion.
