How to Use AI with LinkedIn Sales Navigator

The safe way to use AI with LinkedIn Sales Navigator is to keep AI outside it. You search, save lists and read profiles in Sales Navigator yourself, type short notes in your own words, then give AI those notes plus public company information. AI helps you plan searches, map the buying committee by role, find an angle and draft connection notes. You send every message by hand.

A note on features: LinkedIn adds AI features to its products and changes them often, and plans differ. Check LinkedIn's current help pages for what your seat includes. This guide describes Sales Navigator only in general terms.

What AI can and can't do with Sales Navigator

In general terms, Sales Navigator lets you search for leads and accounts with filters, save them to lists, get alerts when something changes and read profiles and company pages. That's research. AI is good at the thinking around it: what to search for, what a change means and what to say.

A CRM is different. With HubSpot you can connect an approved assistant or export your own records (see How to Use AI with HubSpot for Sales). LinkedIn's terms limit how member data can be copied and used, so the safe path here is manual: your eyes, your notes, your hands on the send button.

Some Sales Navigator plans can connect to a CRM. What syncs, and what you may do with it, is set by LinkedIn's terms and your admin, so check LinkedIn's current documentation and ask RevOps before treating any synced LinkedIn data as something you can put into AI.

Set up before you start

You need three things: your company approved AI assistant with your seller context file loaded (how to build one), a clear answer from your manager or policy on what LinkedIn data may go into it, and a simple notes template. The template does the job a field guide does for a CRM: it tells AI what your notes mean, and it keeps you from copying profiles.

# Account notes: {company}
Why this account (fit, in my words): {one line}
Public sources I opened: {company site, press release, job post, with dates}
Trigger, if any: {what happened, when, where I saw it}

# People I found and confirmed myself
Role in deal | Title | Time in role | One business fact | Where I saw it
{champion?} | {title} | {about 1 year} | {posted about onboarding} | {their profile, date}

# Rules for AI
These are my own notes, typed by me. Use only these and the public sources listed.
Never add personal details. If something is missing, write MISSING.
Mark anything I haven't confirmed as UNVERIFIED.

Workflow 1: Turn your best fit customer profile into a search plan

Most weak prospect lists come from weak searches: one title, no exclusions, no disqualifiers. AI is good at turning your best fit customer profile and personas into a fuller search plan. You then match each criterion to the filters you actually have.

Using the best fit customer profile and personas in my seller context file, build a Sales Navigator search plan for VP of Sales and Sales Enablement leaders at US logistics and supply chain software companies with 200 to 1,500 employees that are hiring prospecting reps.

Give me:
1. Title variants to include, and similar titles to exclude
2. Seniority and function, in plain words
3. Company size and industry criteria
4. 10 keywords to include and 10 to exclude
5. 3 disqualifiers that mean I should skip an account

Describe each criterion in plain words. I will match it to the filters I have. Don't name filters or product features.

Why plain words: a model will happily invent a filter that doesn't exist. Asking for criteria, not menu steps, keeps the plan useful whatever your seat includes.

Run the search yourself and save it. If the results look wrong, tell AI what came back in one line ("too many recruiters, too few enablement titles") and ask for a revised plan.

Workflow 2: Map the buying committee by role

Single threaded deals stall when your one contact goes quiet. AI tells you who to look for; Sales Navigator helps you find them. Always in that order: AI lists roles, you find the people.

We sell Ramplight, AI coaching that helps new sales reps ramp faster, to logistics companies with 200 to 1,500 employees. List the roles usually involved in a purchase like this at Brightwater Cold Chain (a fictional company): economic buyer, champion, users, technical or security reviewers and likely blockers.

For each role: likely titles, what they care about, what would make them say no and one question to ask them.

Roles only. Do not name or guess any people.

Now find the real people for each role in Sales Navigator and your CRM, confirm each current title and add a line to your notes. Then let AI check your coverage.

Here is the committee map for Brightwater Cold Chain and my notes on the people I found and confirmed myself (role, title, time in role, one business fact each).

1. Mark each role as found, not found or unconfirmed
2. Tell me where I'm single threaded
3. Name the missing role that matters most for this deal, and why
4. Tell me who to reach first and why

Use only my notes. Don't add people or details.

<my_notes>
{paste your notes template}
</my_notes>

Workflow 3: Triage your alerts and list changes

Saved lists and alerts can flag job changes, company news and posts. Some are real reasons to reach out. Most are noise. Read them yourself, then type one line per item in your own words and let AI help you sort them.

Below are my one line notes on this week's changes in my saved Sales Navigator lists, typed in my own words, plus public sources I opened.

Sort them into three groups: real trigger worth outreach this week, worth watching, and noise.

For each real trigger: which of our use cases it connects to, a one sentence angle and the next action (connection note, email, call or nothing yet). Mark anything I still need to verify as UNVERIFIED. Use only my notes and the sources I list.

<my_notes>
1. Brightwater Cold Chain: new VP of Sales started last month (her profile)
2. Brightwater Cold Chain: 9 prospecting job posts open (careers page, today)
3. {account}: {change, in your words} ({where you saw it})
</my_notes>

Workflow 4: Draft messages by role, send them yourself

Once you have a confirmed committee and an angle, AI can draft a different reason to connect for each role, all telling the same account story. That's how you multithread without sending three copies of one note.

Using my Brightwater Cold Chain notes and the angle "a large class of new reps starting next quarter while the enablement team is new," draft one LinkedIn connection note for each of: VP of Sales, Sales Enablement Manager, CFO.

Rules: same account story, a different business reason for each role. Under 300 characters each, with the count shown. No pitch, no link, no meeting ask, no product name. Never imply we've met. Use only facts in my notes.

I will edit and send each one myself.

The 300 character rule is our guidance from the course. LinkedIn sets its own limits, which have changed over time, so check its current help pages. For the full method on connection notes, first messages, comments, follow ups and InMail, read How to Write LinkedIn Messages with AI.

1. Run the 60 second edit

   Verify every fact, delete any warm up, read it aloud.

2. Recheck the role

   if your notes are more than a week old.

3. Send it by hand,

   one person at a time, and log the touch in your CRM.

Workflow 5: Tier your saved accounts

Tiering tells you where to spend 15 minutes per account and where 1 minute is enough. It needs account level facts only: public firmographics (industry, size, region) and your one line notes.

Here are 20 accounts from my saved list, with public firmographics and my one line notes. Score each from 1 to 10 against the best fit customer profile in my seller context file, with a one line reason.

Group them into top priority (fully custom outreach), second priority (persona plus trigger) and third priority (segment templates). Flag any account where the score depends on something I haven't verified.

Spot check five scores yourself. Models sometimes weight criteria oddly. Once the prompt works, save it as a prompt template so you reuse the same version each month.

Worked example: Brightwater Cold Chain

What not to automate on LinkedIn

Some tools, including some AI sales agent products, promise to prospect on LinkedIn by themselves. Read LinkedIn's current User Agreement and your company policy first. Most reps should simply say no.

For a broader view of what to automate in sales, and how to judge AI sales agents, read AI Sales Automation.

Rolling it out to a team

These workflows run on notes, not connectors, so they're easy to share.

1. Write the rule down

   Your manager or RevOps confirms, in writing, what LinkedIn data may go into the approved assistant. Start from the "fine" and "never" table above.

2. Share one notes template and the six prompts

   in a team workspace. One owner updates them; reps don't keep private copies.

3. Pilot with three reps for two weeks

   on Workflows 2 and 3. Track how many accounts move from single threaded to multithreaded.

4. Review drafts together weekly

   Read five sent notes aloud and agree on what to fix in the shared prompts.

The AI for Sales Certification teaches these research and outreach habits on reps' own accounts. Sales leaders can see team options. To pair this with your CRM, see How to Use AI with HubSpot for Sales or How to Use AI with Salesforce for Sales.

Frequently asked questions