How to train your team to use AI: a practical plan for the sales team
Training your team to use AI means teaching each role to apply the tool to the tasks it already does, with rules and review. It works better in short cycles with real cases than in a one-off course. Start with marketing and sales, where the gain shows up fastest in revenue.
The common mistake is hiring a talk about the future of AI, leaving excited and changing nothing on Monday. Knowing what a language model is doesn't help a sales rep prepare a better meeting. What helps is an exercise with their actual meeting, a ready-made prompt template and someone who checks the result. This guide shows how to organize that in four weeks.
What does the team really need to learn?
Four skills cover most day-to-day use:
| Skill | What it means in practice |
|---|---|
| Asking well | Giving context, goal, format and an example |
| Checking | Verifying facts, numbers and tone before using |
| Protecting | Knowing which data can't go in |
| Iterating | Correcting the output instead of accepting the first one |
None of them requires technical knowledge. They require practice with work situations. And the second one matters more than it seems. A model can give a wrong answer with complete confidence, and anyone who hasn't learned to be skeptical will pass the error on to the customer.
How do you choose what to train first?
Before any class, build a task map. Gather the team for 40 minutes and ask each person to list what they do every week, how long it takes and how much they hate doing it. Repetitive, time-consuming, low-risk tasks are the starting point.
Examples by role:
| Role | Tasks to train first | Where the gain shows up |
|---|---|---|
| SDR | Account research, personalized first message | Replies and meetings booked |
| Sales rep | Preparing meetings, summarizing, writing follow-ups | Selling time, stage progression |
| Sales manager | Analyzing the pipeline, preparing 1:1s | Quality of coaching |
| Marketing | Ad variations, content ideas, campaign analysis | Testing speed |
| Customer service | Drafting replies, summarizing cases | Response time |
Choose one task per role. Trying everything at once creates curiosity, not habit.
How do you structure the four-week plan?
Week 1: foundations and rules
Run a 60-minute session. Show what the tool does well, where it fails and what the company's rules are. If there are no rules yet, first write the AI usage policy. Close with a real exercise: each person uses the tool on a task from their own week and shares the result.
Week 2: main task by role
Split by role and work with real cases. A sales rep brings a meeting from the following week. A marketer brings a campaign in progress. The internal instructor shows the prompt, the result and what was corrected. Seeing an error get corrected teaches more than seeing a perfect result.
Week 3: practice and library
Each person uses the tool on the chosen task every day. At the end of the week, gather the best prompts in a shared library, with an example input and output.
Week 4: measurement and next cycle
Compare the task before and after. Keep what worked, drop what didn't and choose the next task. From here on, a 30-minute biweekly session usually sustains the pace.
Take Nexo's free assessment and see which of your sales team's tasks eat up the most selling time.
Take the free assessment
How do you teach people to ask well?
A simple five-part template solves most cases: role, context, task, format and limits. Compare two prompts for preparing a meeting:
Weak: "Help me prepare a sales meeting."
Better:
You're a B2B consultative sales rep. I'm going to talk to the operations director of a midsize distributor. He said he spends too much time checking orders. We sell a [description] service. List 7 discovery questions, 3 likely objections with responses and one next step to propose at the end. Answer in short bullet points. Don't make up data about the company.
The second version includes the role, scenario, pain point, format and limit. Note that the company's name was left out, because it isn't needed. The final request, not to make things up, reduces the risk but doesn't replace checking.
Practical examples for sales and marketing
Summary and follow-up
After an authorized meeting, the sales rep pastes their own notes and asks: "Summarize this into decisions, open items and next step, and write a follow-up message of up to 80 words." The rep reviews it, sends it and updates the CRM. The task now gets done the same day, which is usually the real difference. See how to create an effective follow-up.
Objection simulation
The sales rep asks: "Act as a skeptical buyer. Raise the objection that it's too expensive and react to my answers." Then they ask for feedback on what could be improved. The exercise complements role-plays between people. See role-play in sales.
Ad variations
Marketing provides the offer, the audience and the real objections gathered from sales, and asks for ten headline variations, each with a different angle. Then it picks three to test. AI increases the number of tests; the criteria still belong to the people who know the customer.
1:1 preparation
The manager uses aggregated pipeline data and asks for patterns, questions and a focus for the conversation with each sales rep. The decision about the person is still theirs.
How do you build a prompt library?
A simple library keeps everyone from starting from scratch. For each item, record:
- task and role;
- ready-made prompt, with blanks to fill in;
- an example of a good result;
- what to always check;
- owner of the item.
Start with a few items that actually get used. A spreadsheet or shared document does the job. At the biweekly meeting, ask someone to show what improved. Recognize those who share, because that builds a culture of learning from each other.
How do you deal with resistance and overconfidence?
Two profiles show up in almost every team, and they call for different responses.
People who resist are afraid of losing their jobs or looking incompetent. Be clear about the intent, for example: freeing up time for more conversations with customers. Start with the task the person hates most. Pairing them with a colleague usually works better than an order.
People who trust too much send without reading. Show examples of real errors, require checking and use the rule: whoever sends is accountable. In sales, a wrong price on a proposal costs more than the time saved.
How do you measure whether the training worked?
Class attendance isn't a result. Measure on three levels:
| Level | Question | How to measure |
|---|---|---|
| Usage | Is the team using it? | People using it on the defined task each week |
| Quality | Is the result good? | Sample reviewed by the manager |
| Impact | Did the number change? | Time on the task, on-time follow-ups, marketing tests, conversion |
Before starting, record how long the task takes. Ask a few people to measure it for a week. Then compare and decide where the freed-up time goes. If the hour saved doesn't turn into more conversations with customers, the gain is lost. To calculate it, see how to calculate AI ROI beyond sales.
Who should run the training?
A good option is to pick one or two internal "champions," people who already use the tool well and are respected by the team. Give them dedicated time. If you need outside support, ask for examples using your own operation and don't settle for a generic class. The sales manager needs to take part: when they expect the new way of working, the habit sticks.
How much time should you set aside each week?
Set a fixed time slot, even a short one, during the first weeks. Without that space on the calendar, practice loses out to the urgency of closing the month, especially on sales teams.
Common mistakes
- Investing in a one-off event with no follow-up practice.
- Training on the tool instead of the task.
- Giving access without data rules.
- Measuring attendance, not changes in the routine.
- Asking everyone to learn everything at the same time.
Conclusion
Training for AI is, at its core, training to do sales work better. Choose one task per role, practice with real cases, document what works and measure the effect on selling time and results. In a few weeks, AI stops being a curiosity and becomes part of the routine.
Want to know where your sales team would gain the most from AI? Take Nexo's free assessment.
Take the free assessment
To strengthen your team's sales foundation, check out sales training and read about training sales reps at scale.
Frequently Asked Questions
Where do you start training the team on AI?
How long does the training take?
How do you deal with people who resist AI?
How do you know whether the training worked?
Founder of Nexo
20+ years transforming sales. Creator of the Método BD, Sistema Varejo Pró and Engrenagem comercial methodologies.