How to Get AI Training When Your Employer Won't

Your employer talks about AI constantly and has trained you on none of it. You’re not imagining the gap. Surveys through 2026 keep finding that only about a quarter of workers have received any real instruction on how to work with AI, while PwC’s analysis suggests the large majority of workers will need to reskill as AI reshapes their roles. The training most people need is not coming from the org chart. So you build it yourself, on a path that’s free, structured, and doesn’t depend on anyone signing off.

Step 1: Pick the skill from your actual job, not a trend list

Skip “what AI skills are hot.” Start from a task you already do that’s tedious, repeatable, or slow. Drafting reports, cleaning spreadsheets, summarizing calls, answering routine emails. That task is your training ground, because you already understand it well enough to judge whether the AI did it right, which is the entire skill.

Step 2: Use the official, free documentation first

The companies that make these tools publish genuinely good guides, and they cost nothing. OpenAI, Anthropic, Google, and Microsoft all maintain free help centers and prompting guides for their consumer tools. Start there rather than with a paid course, because the documentation is current and the courses often aren’t. An afternoon with the official guide for one tool gets you further than a twelve-hour video series covering tools you’ll never open.

Step 3: Train on real work, in short reps

Don’t set aside a “learning weekend.” Instead, take one real task this week and do it twice, once your normal way, once with the AI tool, and compare. The comparison is the lesson: you’ll see exactly where the tool helped, where it produced confident nonsense, and what you had to add. Repeat with the same task until the AI version is reliably faster, then move to a new task. Skill builds from reps on real work, not from passive consumption.

Step 4: Use free courses as scaffolding, not as the goal

There are good free structured courses, university and vendor “AI fundamentals” programs, and some employers and public agencies now fund reskilling you can tap even if your direct manager won’t. Use them to give your self-training a spine: an order to learn things in, vocabulary, a sense of what you don’t know.

But hold the line on what they’re for. A completed course is not the same as a demonstrated skill, and collecting them becomes its own procrastination, a point we argue in full in Stop Collecting Certificates. The course is the scaffolding. The thing you build with it, the task you can now do better, is the actual result. If you finish a course and can’t point to work you do differently because of it, you learned about AI instead of learning to use it.

Step 5: Make your training visible

The reason to do this isn’t just competence, it’s that self-taught AI skill is invisible until you surface it. Once a tool genuinely improves a task, tell your manager what changed and show the before-and-after. That does two things: it puts you on record as the person who improves how work gets done, and it sometimes shakes loose the budget for the formal training that “wasn’t available” right up until someone proved it pays off.

The training you give yourself counts double once someone else can see the result.

You don’t need permission or a budget to start, and waiting for either is the most expensive option on the table. The quarter of workers getting trained aren’t smarter than you. Many of them just work somewhere that decided to invest. You can decide to invest in yourself this week, for free, starting with one task.