AI for sales: 10 practical use cases, and what to check before you hit send
By Bodhih Training Solutions · UpdatedThe short answer
The most useful applications of AI in sales are account research, list building, outreach drafts, call preparation, call notes, proposals, business cases, objection practice, pipeline analysis and automating one repeatable weekly task. In every case the AI produces a first draft and the seller remains responsible for it. Before anything reaches a buyer, check the facts, the figures, the promises and whether you have permission to contact that person.
What can AI actually do for a salesperson?
A general-purpose assistant such as Claude, ChatGPT, Copilot or Gemini is good at three things a seller does all week: reading a lot of material quickly, producing a structured first draft, and playing a role so you can rehearse. It is poor at knowing what is true about your customer, what your company has approved, and what you have permission to do.
That split decides how to use it. Give the assistant the reading and the drafting. Keep the judgement, the facts and the commitments for yourself. The ten use cases below follow the selling week in order, and each comes with the check that stops a fast draft becoming an embarrassing email.
What are the 10 practical use cases for AI in sales?
The table lists each task, what the assistant contributes and what a human must verify before the work is used.
| Sales task | How AI helps | What a human must check |
|---|---|---|
| 1. Account research | Summarises public information into a one-page brief with likely priorities. | Open every source. Mark each fact as sourced or inferred, and check the dates. |
| 2. Trigger events and stakeholder mapping | Suggests who is in the buying group and what recent change may create a need. | Names and titles go out of date. Confirm people are still in the role. |
| 3. Ideal customer profile and lead lists | Turns a loose description of a good customer into explicit rules for filtering a list. | Count a sample by hand. Confirm the rules select the companies you meant. |
| 4. Outreach sequences | Drafts a multi-touch sequence for each buyer role, with varied angles. | Consent and do-not-disturb preferences, the tone, and that every claim is approved. |
| 5. Call preparation | Builds a call plan and a set of discovery questions from your account brief. | Remove questions the buyer has already answered. Keep the plan to one page. |
| 6. Call notes and follow-up | Turns a transcript into notes, next steps, a CRM update and a follow-up email. | Compare the notes with the transcript. Confirm the call was recorded with consent. |
| 7. Proposals | Reorders your standard proposal around this buyer’s stated priorities. | Every promise, date and scope item must be something your company agreed to. |
| 8. Business case and ROI | Structures the payback argument and drafts the summary for a finance reader. | Recompute the key figures yourself. Use the buyer’s numbers, not assumed ones. |
| 9. Objections and negotiation | Collects objections from past deal notes and plays the buyer so you can rehearse. | Your walk-away point and what you may trade are your decisions, not the AI’s. |
| 10. Pipeline review and forecasting | Analyses a CRM export for stalled deals, slipping dates and thin coverage. | Check the metric definitions and recount two numbers before you present them. |
How do you brief AI so the output sounds like you and not like everyone else?
Most weak output comes from a weak brief. “Write a cold email to a CFO” gives the assistant nothing to work with, so it falls back on the average of every sales email it has seen. A senior colleague given the same instruction would ask you questions first. Answer those questions in the prompt.
- Role: who the assistant is writing as, and at what level of seniority.
- Audience: the buyer’s role, what they care about and what they already know.
- Goal: the one action you want from the reader.
- Context: the account facts, the trigger event and where the conversation stands.
- Examples: one or two pieces of your own writing that worked.
- Format: length, structure and what to leave out.
What should you check before you hit send?
Anything an assistant drafts goes out under your name, and a written promise in an email or proposal can bind your company. Run the same short check every time, whatever the task.
- Facts: can you point to a source for every statement about the buyer?
- Figures: have you recomputed the numbers that matter?
- Promises: is every commitment on price, scope, dates and outcomes approved?
- Proof: are the customer examples real and cleared for use?
- Permission: are you allowed to contact this person on this channel?
- Competitors: is every comparison fair and accurate?
- Tone: would you say this sentence aloud to the buyer?
What should you never paste into an AI tool?
Treat an assistant like any other outside system. Before pasting, ask whether your company has approved the tool for customer information, and whether the buyer would be comfortable seeing their details there. Business contact details are still personal data, and call recordings need the other party’s consent.
When in doubt, remove names and identifying details, or work with a description of the situation instead of the document itself. Your organisation’s data policy and compliance team are the authority here. This guide is general professional guidance, not legal advice.
Which sales tasks are worth automating first?
Automation suits a task that you repeat weekly, that has clear inputs, and where a mistake is caught before it reaches a customer. A pipeline review pack, a call-preparation brief or a first-draft follow-up are sensible starting points. Sending messages to buyers without a person reading them is not.
Start with one task. Write the steps down as you would for a new colleague, run the workflow twice on real work, and add an approval gate at every point where something leaves your hands. Keep a short log of what went wrong. That log is what turns an experiment into a method your team can trust.
How do you turn these use cases into a habit?
Pick the two tasks that cost you the most time this week and save the prompts that worked. A personal prompt playbook of ten briefs covers most of a selling week. Then measure something simple, such as preparation time per call or replies per sequence, so you know whether the change is real.
If you want structure and an outside check on your work, Bodhih’s AI for Sales course covers these ten use cases across eight modules, with each one practised on your own accounts and graded by a human reviewer.
Practise all ten on your own accounts, and have the work graded
Bodhih’s AI for Sales course takes about 16 hours online. You build eight work products on your own accounts, a human reviewer grades them, and you sit a proctored exam for a verifiable credential. It costs ₹7,999 plus GST for an individual.
Questions people ask next
Will AI replace salespeople?
AI is unlikely to replace salespeople in complex or high-value selling, because buyers still want a person who understands their situation and stands behind a commitment. What changes is the mix of work. Research, drafting and analysis become faster, which leaves more time for conversations. Sellers who can brief an assistant well and check its output have an advantage over those who cannot.
Can AI write cold emails that get replies?
AI can write a good first draft of a cold email when you give it a specific brief: the buyer’s role, a real trigger event, one clear ask and an example of your own writing. Without that, the draft reads like every other AI-written email. Always edit for tone, verify every fact about the account, and confirm you are permitted to contact the person.
Is it safe to put CRM data into an AI assistant?
It is safe to put CRM data into an AI assistant only when your organisation has approved that tool for customer information. Check your company’s data policy first. If the tool is not approved, remove names and identifying details or use a made-up sample with the same structure. Business contact details count as personal data, so treat them with the same care as any customer record.
How accurate is AI at sales forecasting?
AI is useful for sales forecasting as an analyst, not as an oracle. Given a CRM export, an assistant can flag stalled deals, slipped close dates and weak coverage quickly. It can also miscount rows or apply a metric differently from your team. Agree the definitions first, recount at least two numbers by hand, and treat the forecast call as your own judgement.
How do I start using AI in sales if my company has no policy yet?
Start with tasks that use only public information and your own writing, such as researching an account from its website or improving a call plan. Avoid pasting customer data, pricing or contracts until your company confirms which tools are approved. Ask your manager or IT team for a decision, and suggest a short written rule the whole sales team can follow.