Generative AI for Sales: Lead Gen, Prompts and Outreach

Generative AI for sales helps a small business reply to every lead fast, qualify it, book the call and follow up without things slipping, while you keep control of the conversations that close deals. In practice that means AI chat and SMS assistants that answer new enquiries, ChatGPT-style tools that research and draft, and automations that log everything in your CRM. Used carelessly, it also means spammy outreach that breaks the law, so this guide covers the rules too.

Here’s how to use AI for sales and lead generation, which tasks to hand over, prompts that work, and how to automate outreach without breaking CAN-SPAM or Canada’s CASL.

How to use generative AI for sales: where it helps

For most service businesses, sales isn’t a big outbound team. It’s an owner or a couple of people juggling enquiries between jobs. The leak is usually in the first hour: a form comes in, nobody replies until tomorrow, and the customer has already booked someone else.

AI helps most at these points:

• Capture. Answer chat, forms, texts and missed calls instantly, any hour.
• Qualify. Ask the questions you’d ask (budget, location, timeline, type of job) and score the lead.
• Book. Offer real calendar slots and send reminders.
• Follow up. Draft the follow-up email or text, personalised from what the lead said, for you to approve.
• Admin. Summarise calls, update the CRM, write proposals from templates.

Closing stays with you. Pricing, negotiation and trust-building are where people still win.

AI lead generation: inbound first

“AI lead generation” gets sold as tools that find thousands of prospects for you. For a local or service business, the better return almost always comes from converting the leads you already get.

Inbound AI lead generation:

• A website chatbot that answers questions and asks for contact details at the right moment (see best chatbots for small business)
• An SMS reply to every form fill within a minute
• An AI answering service that picks up missed calls and books jobs
• Automated lead routing so the right person gets notified immediately

Outbound AI lead generation:

• Building target lists from public sources and data tools (Apollo, LinkedIn Sales Navigator, industry directories)
• Enriching records with company size, website and role
• Researching each prospect before you write to them
• Drafting first-touch emails that reference something real about the prospect

Real estate is a good example. AI for real estate leads typically means instant replies to portal enquiries, a few qualifying questions (buying or selling, area, price range, pre-approved?), and a booked call or showing. We cover the details in AI for real estate.

AI appointment setters

An AI appointment setter replies to a lead, qualifies them and books a meeting on your calendar. There are three kinds:

On voice: in February 2024 the FCC ruled that AI-generated voices count as “artificial” voices under the Telephone Consumer Protection Act, so outbound AI calls to cell phones need the same prior consent as robocalls (Cooley summary of the FCC ruling). AI for cold calling, in the sense of an AI dialling strangers, is a legal minefield. AI that calls back people who asked you to call is fine.

ChatGPT for sales: prompts that work

ChatGPT, Claude and similar assistants are useful sales assistants if you give them context. A business plan keeps your data out of model training; ChatGPT Business, for example, is $20 per user per month billed annually as of November 2026 (OpenAI help center).

Prospect research

Here is the public website text of [company]. Summarise in 5 bullets what they do, who their customers are, and 2 likely operational problems a [your service] provider could solve. Flag anything you’re guessing.

First-touch email

Write a 90-word email to [name], [role] at [company]. Reference [specific thing from their site or a recent post]. Offer [one specific outcome]. One question as the call to action. No flattery, no “I hope this email finds you well”.

Follow-up after no reply

Write a 50-word follow-up to this email: [paste]. Add one new useful point, not a “just checking in”.

Objection prep

I’m selling [service] at [price] to [type of customer]. List the 6 most likely objections and a short, honest answer to each.

Call summary to CRM

From this call transcript, give me: budget, decision maker, timeline, main problem, next step with date, and a 2-sentence summary for the CRM. Transcript: [paste].

Proposal draft

Using this template [paste] and these call notes [paste], draft a proposal. Keep my pricing exactly as written. Mark anything you’re unsure about with [CHECK].

Always read before sending. AI will invent details about a prospect if you let it, and nothing kills a cold email faster than a wrong fact about the reader’s business.

How to automate cold outreach responsibly

Cold email can work for B2B services. It also gets businesses blacklisted and fined when done at scale without care. Here are the rules that matter for US and Canadian senders.

United States: CAN-SPAM

The FTC’s CAN-SPAM compliance guide sets out the rules for commercial email, including B2B:

• Honest “From”, “Reply-To” and subject lines
• Identify the message as an ad where required
• Include a valid physical postal address
• Give a clear way to opt out, and honor it within 10 business days
• You stay responsible even if an agency or tool sends for you

Penalties are up to $53,088 per violating email, per the FTC guide (checked November 2026).

Canada: CASL

Canada’s Anti-Spam Legislation is stricter. It generally requires consent before you send a commercial electronic message, express (they said yes) or implied. Implied consent from an existing business relationship typically lasts two years from a purchase, and the window after a simple enquiry is much shorter. Every message must identify you, give contact details and include an unsubscribe that works within 10 business days. Maximum penalties are $10 million per violation for businesses and $1 million for individuals.

In practice: don’t run cold email campaigns to Canadian addresses scraped from the web. Use inbound, referrals, events and conversations that create consent.

LinkedIn automation

People search “how to automate LinkedIn outreach” a lot. LinkedIn’s rules on prohibited software ban bots and browser extensions that automate actions like connection requests and messages, and accounts that use them risk restriction. Use AI to research and draft; send by hand.

A responsible outreach setup

1. Send from a separate domain set up for outreach, with SPF, DKIM and DMARC configured.
2. Keep volumes low per inbox and warm new inboxes up gradually.
3. Verify addresses before sending to keep bounces low.
4. Personalise from real research, not a merge field and a compliment.
5. Stop the sequence the moment someone replies, and honor unsubscribes everywhere.
6. Log every send and opt-out in your CRM.

AI tools for sales reps

Group tools by job rather than chasing a “best AI tool” list:

Sales workflow automation (moving a lead from form to CRM to calendar to follow-up) is usually where the hours are saved. We show specific recipes in CRM workflow automation and broader ones in marketing automation for small business.

A simple AI sales setup for a service business

1. Website chat and form, both replying within a minute (chat via a bot, forms via SMS or email).
2. Qualifying questions that match how you price jobs.
3. A booking link with your real availability.
4. Every lead, answer and booking written into your CRM automatically.
5. AI-drafted follow-ups at day 1, day 3 and day 7, approved by you.
6. A weekly review of lost leads: why did they go cold?

You can build most of this with off-the-shelf tools and a Zapier, Make or n8n account. If you’d rather have it set up and tested for you, that’s the kind of work opsmigo does for service businesses.