
Where AI helps in prospecting (and where it doesn't)
If you are wondering how to use AI for sales prospecting, start with one rule: use it to do the research and the first draft, and keep the judgment and the relationship for yourself. AI is good at reading a lot quickly, spotting patterns and producing a starting point. It is bad at knowing which of those patterns matter to a specific buyer this week.
| Task | Good fit for AI | Keep human |
|---|---|---|
| Account research | Summarize a company's site, news and job posts | Decide if the account is worth the time |
| List building | Group accounts by traits you define | Set the criteria and check the list |
| First drafts | Draft an email from your notes | Edit for tone and truth |
| Call prep | Pull together past calls and open items | Choose the angle for this call |
| Follow-up | Draft a recap from the transcript | Confirm details and send |
1. Research accounts faster
Before you reach out, you want three things: what the company does, what changed recently and who owns the problem you solve. A general AI assistant can summarize a company's website, recent announcements and open job posts in a minute. Give it a clear prompt and ask it to say when it is unsure.
You are helping me prepare to contact [Company] ([website]). Using only the sources I paste below, summarize: 1. What they sell and to whom, in two sentences. 2. Any recent changes (new leaders, funding, launches, hiring) with dates. 3. Which team is most likely to own [problem I solve], and why. 4. Two questions I could ask on a first call. If a point is not in the sources, write "not found" instead of guessing. Sources: [paste text from their site, press page, job posts]
The last line matters. AI tools can state guesses as facts. Asking for "not found" and pasting your own sources keeps the research grounded. Always check anything you plan to mention to a prospect.
2. Build better target lists
AI can help you sort and group accounts once you have defined what a good fit looks like. Write down your ideal customer in plain terms first: industry, size, team structure, the trigger events that make them ready to buy. Then use AI to tag a list against those traits and explain each tag.
- Look at your wins first. What did the deals you won have in common? If you record calls, the reasons buyers gave in their own words are the best input you have.
- Use triggers, not just firmographics. "Hiring five or more reps" is a better signal than "50 to 200 employees."
- Keep a human check. Review a sample of every AI-tagged list before you load it into a sequence.
3. Write first drafts, not final emails
AI-written outreach is easy to spot when nobody edits it: long, polite, vague, and the same as everyone else's. Use AI for the skeleton, then make it yours.
Write a cold email under 90 words to [role] at [Company]. Reason for writing: [specific trigger]. Problem I think they have: [problem], because [evidence]. Ask: a 20-minute call next week. Tone: plain, direct, no hype words, no exclamation marks. Do not invent facts, numbers or customer names.
Then edit: cut the first sentence (it is usually filler), add one detail only you would know, and read it aloud. If it sounds like a robot, it is. Our list of sales email subject lines pairs well with this step.
4. Prepare for calls with AI briefs
Prospecting does not stop at the first reply. The first call is where most deals are won or lost, and preparation is where AI saves the most time. A good pre-call brief answers: who is on the call, what happened last time, what is still open, and what you want to leave with.
Call: Discovery, Fernway Logistics, Thu 10:00 Who: Morgan Lee (Head of Sales), first conversation Context: Hiring 6 SDRs this quarter. Said on the cold call: "I can't listen to everything." Open items: none yet Goal for this call: understand how coaching works today and who signs off on tools Questions to ask: - How do you pick which calls to review? - What does a new rep's first month look like? - Who else would be involved in choosing a tool?
Example only. Fernway Logistics and Morgan Lee are made up for this guide.
LuminaNotes builds pre-call briefs like this automatically from your past calls, open action items and CRM context, so you walk in knowing what was said last time.
5. Follow up while the call is fresh
The fastest prospecting win is often the follow-up you already owe. After a first call, an AI note taker can turn the transcript into a summary with action items and draft the recap email, so it goes out the same day. See our follow-up email templates for the structure, and automatic recap emails for how LuminaNotes sends them.
6. Learn from the calls you already have
The richest prospecting data most teams own is sitting in their call recordings. Buyers tell you why they took the meeting, what almost stopped them and which words they use for their problem. That is better input for targeting and messaging than any generic list.
- Ask questions across your calls. "Which competitors came up most last quarter?" or "What reasons did buyers give for taking the first meeting?" With LuminaNotes, you can ask your meetings in plain English and get an answer that points to the calls it came from.
- Look at win and loss patterns. Compare the first calls of deals you won and lost. Which questions, objections and next steps show up in each?
- Reuse the buyer's words. If three prospects described the problem as "chasing updates", put that phrase in your next subject line.
Mistakes to avoid
- Sending unedited AI emails at scale. More volume of generic email rarely means more meetings, and it can hurt your sender reputation.
- Trusting AI research without checking. Mentioning a "recent launch" that never happened ends the conversation.
- Pasting private data into tools you have not vetted. Check your company's policy before you paste customer information into any AI tool.
- Skipping the human part. Buyers still buy from people. Use the time AI saves to have better conversations.
A simple AI prospecting workflow
- Define your ideal customer and three trigger events in a shared doc.
- Use AI to research and tag a list, then review a sample by hand.
- Draft outreach with AI, edit every email, and send in small batches.
- Record first calls with an AI note taker and get a summary with action items.
- Use BANT or MEDDIC scoring after each call to decide where to spend your time.
- Send the recap the same day and keep your CRM updated with two-way sync.