Checklist for AI outreach in a new market

A practical guide for founders and SDRs to validate a new B2B market on LinkedIn using AI lead scoring, intent classification, and iterative messaging.

Get started

Short answer

Testing a new market on LinkedIn means running a small, separate test: write down who you think feels the pain and why now, send a message that asks about the problem instead of pitching, and sort replies by intent. Then change one thing at a time until the replies tell you whether the problem is the targeting or the script.

Define the new buyer profile

When you enter a new vertical, you don't know which job titles actually feel the pain you solve. If you target too broadly, you'll spend your account's daily limits on people who have no budget or authority. Write the profile down before you search, so the search has something to check against.

Start with a trigger event. This is a specific change in a company's status, like a new product launch, a funding round, or a leadership change, that makes your solution a priority today rather than next year. Without a trigger you are competing with every other item on their list.

  1. 1

    Map budget owners

    List the 3 to 5 job titles that can sign a contract. Target the person who pays for the tool, not just the end user.

  2. 2

    Identify the trigger

    Determine the exact event that creates the need. If you sell compliance software, the trigger might be a company expanding into a new geographic region.

  3. 3

    Write your hypothesis

    Draft a one-sentence statement: people with this title at this kind of company care about this pain because of this trigger.

  4. 4

    Build a benchmark list

    Find 20 to 50 LinkedIn profiles that match this description. Use them to check whether your search filters return the right people before you contact anyone.

Build a testing lead group

Don't mix new leads with your proven ICP. If you send an experimental script to your core market, you risk damaging your reputation with your best prospects, and you lose any clean way to tell which results came from the test.

You need a sandbox. Create a dedicated group for the new vertical so you can track acceptance rate and reply quality in isolation. Keep the group small enough that a bad script doesn't cost you the whole segment.

The spray and pray approach

  • Adding 500 leads from three different industries into one campaign
  • Using a generic 'I help businesses grow' script for everyone
  • Measuring success by the number of connection requests sent

The surgical approach

  • Creating a group of 100 leads specifically for fintech VPs
  • Using a script that mentions a specific fintech regulation change
  • Measuring success by the number of positive intent replies

Limit your first test group to 100 to 200 people. This prevents you from burning through an entire segment with a script that doesn't work.

Draft a low-friction hypothesis

In a new market, the first message is a question about the problem, not a pitch. If you ask for a meeting in the first message, you'll get ignored because there's no trust yet and they have no reason to believe you understand their week.

The first message should seek a yes to the problem. Once they agree the problem exists, the meeting becomes a logical next step. If they say no, you've learned something about the trigger or the title, and you can change one of those instead of rewriting everything.

Opening

Element
I see you are the VP of X and I have a solution for Y
Avoid this
I noticed [specific event] and wondered if [pain] is a priority

Value prop

Element
We help companies grow faster with our platform
Avoid this
Most [titles] I talk to are stuck on [specific problem]

Call to action

Element
Do you have 15 minutes on Tuesday for a demo?
Avoid this
Is this something you're currently focused on?

Paste this job

Connection request template (under 200 characters):

Hi [Name], noticed [Company] is expanding into [Region]. Usually that makes [pain point] a headache for [title]s. Is that a priority for you right now?

Classify intent to iterate

Counting replies is a vanity metric. A person saying 'Not interested' is a reply, but it doesn't mean your market is valid. Classify replies by intent to know if you should pivot your script or your targeting.

If you see a pattern of soft intent, your target audience agrees they have the problem but doesn't believe your specific solution is the right way to fix it. Fix the message before you change the targeting.

Write down the exact phrases people use. 'We already have someone for that' and 'we're not big enough for that yet' point to different fixes, and you only see the difference if you keep the words.

  1. 1

    Bucket the replies

    Label every response as positive intent (asking for a demo or pricing), soft intent (asking for more info or a deck), or no interest.

  2. 2

    Analyze the gaps

    High soft intent with low positive intent means your value proposition is too vague. You've identified the pain but not the cure.

  3. 3

    Validate the target

    A high no-interest rate usually means your lead targeting is off or the pain point isn't a priority for that role.

  4. 4

    Mirror the language

    Take the exact words used by positive intent leads and put them into your next script version.

Example

  1. Suggested reply

    Hi Sarah, noticed you're hiring 5 new SDRs. Usually onboarding them takes forever. Is that a priority for you?

  2. Prospect

    It is, but we just bought a tool for this. How is yours different?

  3. You

    Fair question. Ours scores each lead so your new SDRs don't spend their first month on bad fits. Worth a quick look?

Decide what result ends the test

Pick the number of leads and the number of weeks before you start, and write down what would make you stop. Without a rule, every test looks like it needs two more weeks, and you burn the segment learning nothing you can act on.

The rule should name the signal, not the volume. Sending 200 invites is an activity. Twelve people saying the problem is real and asking what you do about it is a signal.

Few connection accepts

Signal
The note reads like a pitch, or the title is wrong
What it means
Rewrite the note around the trigger event and check the title against your benchmark list

Accepts but no replies

Signal
The first message asks for too much
What it means
Change the ask to a yes or no question about the problem

Replies that agree but don't book

Signal
The pain is real, the offer is unclear
What it means
Name the outcome in one sentence and drop the feature list

Replies that say the problem is handled

Signal
Someone else already owns this, or the trigger isn't live
What it means
Tighten the trigger filter and test a different title

Doing this with Kinlead

You can run this test inside Kinlead. Describe the new buyer in the lead search or paste your website, and each lead gets a score from 0 to 100 against that profile. Leads below the fit line are not kept, so the test group stays clean.

Put the test leads in their own lead group and point a campaign at it. Connection notes are written by AI under 200 characters and open with why that person became a lead. Use hand-off mode for the first week so you read every reply yourself, then switch to one of the AI reply modes once the script holds.

Replies from every campaign land in one inbox, classified by intent with interested replies first, and you get an email when someone replies. If you want to see the language before you write anything, the Steal Customers agent watches competitor pages and turns commenters into leads, keeping the post URL and what they commented.

  • Set the send window and daily limits so the test runs at a human pace.
  • Pause the campaign the moment the signal is clear, either way.
  • Export the test group to CSV if you want to review the list outside Kinlead.

Frequently asked questions

Related

Get started with Kinlead

Prefer markdown? Checklist for AI outreach in a new market.md