AI Consulting Sefa Aydın 7 min read

How to Run an AI Training Needs Analysis: 5 Steps Before Corporate Training

How to Run an AI Training Needs Analysis: 5 Steps Before Corporate Training

Short answer

An AI training needs analysis is a short assessment you run before designing the training. It shows which tasks teams could use AI for, how often and in what ways they use AI tools today and what level they are at. It has five steps: list recurring tasks, see current use through a survey and short conversations, measure the level with a small exercise, write down the rules, then record a baseline and set priorities.

Requests for corporate AI training usually arrive as a single sentence: "We want AI training for our team." Behind that sentence, though, the needs can be very different. On one team nobody has opened the tool yet, on another a few people use it every day without the company knowing, and on a third the licenses have been bought but nobody knows what to use them for. A program designed without seeing this picture first feels too basic for half the room and too fast for the other half. That is why I run a short needs analysis before every corporate training program. In this article I explain what an AI training needs analysis is, how to run it in five steps, which questions the survey should include and what you should have in hand at the end.

What is an AI training needs analysis?

An AI training needs analysis is a short assessment you run before designing a training program. It shows which tasks teams could use AI for, how often and in what ways they use AI tools today, what level they are at and which rules they need to work within. Its purpose is to turn the training from a general tool demo into something tied to the team's real work. A good analysis is not a long report. At the end you have a short task list per team, a skill level breakdown, written rules and baseline measurements to compare against after the training.

Why isn't generic AI training enough?

A generic AI training session shows what the tools can do, but it does not tell participants what to do when they get back to their desks the next day. The result I see most often is this: interest is high on training day, and two weeks later usage has dropped back to where it was before. I cover the reasons in detail in my article on why employees don't use AI. The biggest reason is that people don't know which part of their work to use the tool for. A needs analysis closes exactly this gap, because every example in the training comes from work the team already does.

The second benefit is matching the training to each person's level. When someone who has never used the tool sits next to someone who uses it daily, a program built without an analysis loses both of them. Based on the analysis, I either split groups by level or prepare separate exercises for different levels within the same session.

How do you run a needs analysis in 5 steps?

  1. Start with the work, not the tool. Ask each team manager to list the 5-10 recurring tasks that take up most of the team's week: preparing reports, answering emails, writing up meeting notes, drafting proposals, compiling data and so on. This list is the raw material for the training exercises.
  2. See how people use AI today. Use a short survey to ask who uses which tool, for which task and how often. In addition to the survey, hold fifteen-minute conversations with two or three people from each team. The conversations show what the survey does not: whether employees are processing company data through personal accounts. I cover this unsanctioned use in my article on shadow AI. The goal of the analysis should be to see the real picture, not to blame anyone.
  3. Measure the level with a short exercise. Self-assessment is misleading: someone who has barely used the tool may rate themselves as good, while an experienced user may say average. Instead, give everyone a small task taken from their own work and look at how they write the prompt and how they check the output. This short exercise gives you the level information a survey alone cannot.
  4. Write down the rules and constraints. Which tools does the company have licenses for, which data must never go into a tool, and is there a written usage policy? The training has to fit within these boundaries. I cover the personal data side under Turkey's data protection law (KVKK) separately in my article on KVKK and AI data rules for employees. If there are no rules yet, at least a short list of data rules should be ready before the training.
  5. Record a baseline and set priorities. For a few of the selected tasks, note how long they take today, how many people use the tool regularly and what the output quality is. Then rank the tasks with two questions: how often does this task repeat, and how much easier does AI make it? Without this record you cannot show later whether the training worked. I explain how to make that comparison in my guide to measuring the impact of AI training.

Which questions should the needs analysis survey include?

Keep the survey short. Here are the question areas I use. A survey that takes more than ten minutes gets abandoned halfway, and the most valuable answers, the ones from people who have never used the tools, are lost.

Question areaSample questionWhat it shows
Recurring tasksWhich three tasks take up most of your week?The source of training examples
Current useWhich AI tool have you used in the past month, and for what task?The starting point and unsanctioned use
FrequencyHow often do you open these tools: never, a few times a month, a few times a week, every day?The skill level breakdown
BarriersWhen you don't use AI, or use it rarely, what stops you?The real problem the training must solve
Data awarenessDo you know which information you should never type into an AI tool?The need for rules and risk training
ExpectationsAfter the training, which task would you like to do faster or better?The success criterion

You may want to make the survey anonymous, but always ask which team they are on. Even without names you get a team-level picture, and in an anonymous survey employees are more willing to report unsanctioned use.

How do you determine the skill level?

Three levels are usually enough for corporate training. A finer split makes the program more complicated without adding much to its design.

  • Beginner: Has never used the tool or has tried it a few times. They first need to learn what the tool can and cannot do, how to write basic prompts and what the data rules are. This foundation is the core of what we call AI literacy.
  • Regular user: Uses it a few times a week for simple tasks. They need to learn to write structured prompts that fit their own work and to check the output systematically.
  • Advanced user: Uses it every day and has developed their own prompts. These people are ready to build workflows and set an example for the team. After the training, they are natural candidates for an AI champion role within their team.

What should you have at the end of the analysis?

  • 3-5 prioritized tasks per team, each with a real example to use in the training.
  • A skill level breakdown for each team, with a group or exercise plan based on it.
  • The tools to be used, their license status and written data rules.
  • Baseline measurements for the selected tasks: time, number of regular users and output quality.
  • A plan for what to measure again 30, 60 and 90 days after the training.

These five outputs do most of the work of drafting the training program. At this point it is clear which team starts with which task, what to emphasize at each level and how success will be measured.

Common mistakes

  • Running the analysis only with managers: Managers know the team's work, but they often don't know how the team uses the tools. That is why short conversations with employees should not be skipped.
  • Relying only on self-assessment: A "rate yourself from 1 to 5" question does not give you a level on its own. Always add a short exercise.
  • Starting with a list of tools: "Which tool should we teach?" belongs at the end of the analysis, not the beginning. Work first, tools second.
  • Skipping the baseline: You cannot show that something improved after the training if you did not measure it before.
  • Leaving the analysis as a report: The value of the analysis lies in turning it into a program. Every finding should show up somewhere in the training.

Where should you start?

My suggestion is to start with one team: list the tasks, run the short survey, hold two or three conversations and note the baseline measurements. This small trial shows what will work across the company. To build a program tailored to your team after the analysis, see the corporate AI training page, and for teams at the beginner level, see the AI literacy training program.

Frequently Asked Questions

What is an AI training needs analysis?

It is a short assessment you run before designing corporate AI training. It shows which tasks teams could use AI for, how much they use AI tools today, what level they are at and which rules they need to work within.

How long does a needs analysis take?

For a single team, the task list, a short survey, a few conversations and a small exercise usually take a few days. As the company grows, the time grows with the number of teams, which is why I recommend starting with one team.

How many questions should the survey have?

The survey should take no more than ten minutes. One or two questions each on recurring tasks, current use, frequency, barriers, data awareness and expectations are usually enough.

Is self-assessment enough to determine employees' skill level?

No. Self-assessment can be misleading. Giving each employee a small task from their own work and looking at how they write the prompt and check the output gives far more reliable information about their level.

How does a needs analysis connect to measuring training impact?

The baseline measurements recorded during the needs analysis are the point of comparison for the measurements taken 30, 60 and 90 days after the training. Without a baseline, you cannot show later whether the training worked.

Want to learn to do this with AI?

In individual training sessions, live on your own project, and in corporate training for teams, I teach step by step how to do everything covered in this article with AI. Write in and I will get back to you the same day.

Message on WhatsApp See individual training Explore AI consulting
Sefa Aydın

Sefa Aydın · AI Trainer & Consultant

An AI trainer and consultant who has worked on the Turkey projects of world-famous luxury brands. He teaches, hands-on, how every kind of work is done with AI: design, video, branding and vibe coding.

About Sefa Aydın →   LinkedIn ↗

Related articles

Want to apply this to your own work?

Tell me what you are working on. I reply the same day with next steps.

Your details are used only to reply to you. See the privacy notice.

Message on WhatsApp