Most sales reps now have Claude, ChatGPT, Copilot, or Gemini open in a browser tab. Ask what they use it for and the usual answer is email polish: tightening a follow-up, fixing the tone, cutting a paragraph in half. Good AI training for sales teams moves reps into deal work, the account research, call prep, and forecast notes that decide how a quarter lands.

The upside is large. Gartner surveyed 1,026 B2B sellers in early 2024 and found that sellers who effectively partner with AI tools are 3.7 times more likely to meet quota than those who do not. Most companies have already bought the licenses. Enablement teams now need a way to teach reps to use those tools on live accounts with the team’s own playbook and messaging. They also need a way to see which reps have the skill.

This guide covers how enablement can do that in three parts. Pick four or five deal tasks, teach each one in a short structured lesson, then put an AI ride-along coach next to reps while they prep a real account in Claude. Every part is scored on a rubric, and managers see the results.

Why Reps Stop at Email Polish

Email polish is the easiest AI win because it needs almost no context. A rep pastes a draft, asks for a cleaner version, and gets one back. Deal work asks more of the rep. They have to know what to feed the tool, which sources the company trusts, what the approved positioning says, and how to check the output before it reaches a buyer.

Time is the other pressure. Salesforce’s State of Sales research found that sales reps spend 60% of their time on non-selling tasks. A lot of that time goes to research, notes, and CRM updates, which is exactly the work AI handles well once a rep knows how to direct it.

Reps also carry a heavy load already. The same Gartner survey found that 72% of sellers feel overwhelmed by the number of skills required for their job, and 50% are overwhelmed by the amount of technology needed. Handing reps another login and a lunch-and-learn adds to that load, so the training has to be small and tied to work they already do every week.

What AI Training for Sales Teams Should Cover

Start with the tasks that show up in every deal and that managers already inspect. Five work well for most B2B teams.

• Account research: The rep builds a short brief from approved sources, such as recent news, press releases, hiring trends, and what the CRM shows about past conversations. A good brief names two or three likely priorities and the evidence behind each one.
• Call prep: The rep turns that research into an agenda, a set of discovery questions matched to the buyer’s role, and the one or two outcomes the call needs.
• Follow-ups that use approved messaging: The rep drafts the recap from call notes and pulls positioning and proof points from the team’s approved messaging. That keeps the email in line with what marketing and product signed off on.
• Mutual action plans: The rep drafts a mutual action plan with dates, owners on both sides, and the buyer’s decision steps, then asks the AI to find gaps in it.
• Forecast notes: The rep summarizes deal status, risks, and next steps in the format the manager uses for sales forecasting, with every claim tied to something the buyer said or did.

Write down what good looks like for each task before you build anything. A strong call prep doc for your team might need a hypothesis about the buyer’s priorities, three discovery questions, and a clear ask. Those criteria become the rubric later, so reps practice against the same standard they get scored on.

Teach Each Deal Task in a Short, Structured Lesson

Keep each lesson short enough to finish between calls, and cover one task per lesson. Yoodli is the AI Platform for Human Skills, and its AI tool training builds lessons like these from the material your team already has.

Point Yoodli at the sales playbook, the messaging doc, and the wiki pages on how reps should use Claude. It generates the tutorial, the practice scenarios, and a knowledge check. Teams can also start from a template in the expanded template library, which includes tool training. The AI Tutor walks the rep through the workflow step by step and answers questions along the way. Then it asks the rep to explain the process back, and Yoodli scores that explanation so you know the rep understood the steps.

An account research lesson might cover which sources your team trusts and how to prompt Claude for a brief on a named account. It should also show reps how to check each claim against a source before it goes into the CRM. The guide to training your team on Claude covers the general setup, including the interface walkthrough and the certification rubric. For a sales team, keep every example on deal work.

When the source of truth changes, only the affected modules refresh, and admins approve the changes before reps see them. Sales teams feel this most, because messaging, pricing, and competitive positioning change often.

Put a Ride-Along Coach Next to Reps on a Real Account

A lesson teaches the steps. The skill shows up when a rep does the work on an account they own, and a ride-along coach helps at that moment.

A ride-along is a type of AI roleplay in Yoodli Roleplays. The rep shares their screen inside Claude, or inside ChatGPT, Copilot, Gemini, or the CRM, and preps a real account. The AI sees what the rep types and what the tool returns. An AI coach sits alongside the rep and helps when the rep asks a specific question. The coach never appears as a character in the scene, and it waits for the rep to ask.

Here is how that plays out. A rep preps for a first call with a VP of Sales Enablement at a mid-market software company. They ask Claude for an account brief, then ask the coach whether the brief is ready to use. The coach notes that two of the priorities have no source and suggests checking the company’s recent press releases. Later the rep asks which discovery questions fit an enablement leader, and the coach answers from the team’s own playbook.

Setting one up takes little work. Admins build a ride-along in the Roleplay Builder, and the Tutor template is the recommended starting point. You describe the setup in plain language, the builder drafts it, and you score two to four goals per session. One simple prompt is enough to create the coach, so there is no separate documentation to keep up to date.

Score Deal Work on a Rubric Managers Can See

Every session is scored against the goals you set. For an account research ride-along, the goals might look like this:

• The brief cites a source for every priority it names.
• The rep checks the AI output against the CRM before using it.
• Discovery questions match the buyer’s role and the team’s sales methodology.
• The follow-up uses approved positioning and proof points.

Yoodli Measure shows progress across programs and against rubrics. Managers get dashboards with session playback, transcripts, and score trends, so a 1:1 can start from what the rep actually did in Claude last week. That matters because the same Salesforce research found that 75% of sales reps say they’re more likely to hit their targets with a coach or mentor.

Programs let you sequence the lessons and ride-alongs and track completion across the team. Run the five deal tasks as one program, and a manager can see which reps have finished each step and where scores run low.

Steady practice pays off for sales teams. USB Payments had its sellers practice in Yoodli and saw a 19% increase in meetings booked within 30 days.

Validate Tool Use and Demos, Too

The same setup works for checking rep work inside sales tools. When a rep shares their screen in the CRM or an AI tool, Yoodli can check whether they completed the workflow the way the team teaches it. Examples include logging next steps in the CRM or writing a forecast note in the right format. Enablement can then certify tool skill alongside the talk track.

Demos follow the same pattern. Loopio’s three-person enablement team used the Yoodli AI Tutor to certify reps on a revamped demo workflow, and the company cut demo certification attempts by 67% and certified 60 reps. New hires got the full guided lesson, and experienced reps could skip ahead to practice.

Let Reps Start Call Prep From Claude With MCP

Some reps will want prep to start where they already work. With the Yoodli MCP connection, a rep can ask Yoodli from Claude to prep them for an upcoming call. Yoodli pulls in the context and builds a roleplay for that conversation, so the rep can rehearse the opening and the hard questions before the meeting.

This fits the deal-work habit you are teaching. The rep researches the account in Claude, builds the call plan, and then practices the call itself, all before the buyer joins.

A Four-Week Rollout Plan for Enablement Teams

Roll out in waves, and start with newer reps. NBER researchers studied 5,179 customer support agents and found that access to a generative AI assistant increased issues resolved per hour by 14% on average. That figure included a 34% improvement for novice and low-skilled workers. Support work differs from sales work, but the finding points enablement toward ramping reps as a strong first cohort.

• Week 1: Pick the five deal tasks, write what good looks like for each one, and turn those criteria into rubrics.
• Week 2: Build the account research lesson and one ride-along, then pilot both with a small group of newer reps and one manager.
• Week 3: Adjust the rubric based on the pilot, add call prep and follow-ups, and open the program to the full team.
• Week 4: Add mutual action plans and forecast notes, and review score trends with managers in their 1:1s.

After the first month, keep the program current. When messaging, pricing, or the AI tool changes, refresh the affected lessons and rerun the ride-alongs that depend on them.

Frequently Asked Questions

How can sales reps use AI in their daily work?

Sales reps can use AI for account research, call prep, follow-up emails, mutual action plans, and forecast notes. Deal work like this returns far more value than email polish. Reps get the most from it when they know which sources to trust, use approved messaging, and check every output against the CRM before it reaches a buyer.

What should AI training for sales teams cover?

AI training for sales teams should cover the deal tasks reps repeat every week, such as account research, call prep, follow-ups, mutual action plans, and forecast notes. Each task needs a short lesson, a clear standard for good output, and practice on a real account, scored on a rubric that managers can review.

How do you train sales reps to use Claude?

Train sales reps to use Claude with short lessons built around one deal task at a time, followed by practice on a real account. A ride-along coach can sit with the rep while they work in Claude and answer questions against the team’s playbook. Score each session on two to four goals so managers can see progress.

How do you measure whether sales reps are using AI well?

Measure AI skill with rubric scores on real deal tasks, since logins and usage counts say little about quality. Score whether research is sourced, messaging is approved, and follow-ups match the call. Then review session playback and score trends with each rep in 1:1s, and check whether the work shows up in deal notes and forecast calls.

Start with one deal task and one cohort of reps, then add tasks as scores improve. To see how a ride-along coach would work with your playbook and your AI tools, talk to our sales team.