# How to train AI to handle common sales objections
> Write a sales playbook that tells AI how to answer LinkedIn pushbacks on timing, budget and competitors, and when to hand the conversation to you.
- HTML: https://getkinlead.com/ai-sdr-objection-handling-playbook
- Markdown: https://getkinlead.com/ai-sdr-objection-handling-playbook.md

**Short answer:** To handle objections with AI on LinkedIn, write a sales playbook that maps the pushbacks you actually get to the replies you already use. The AI reads that playbook when it writes a message, so it answers timing, budget and competitor pushback with your logic instead of a generic pitch. Then set a rule for when it stops and hands the chat to you.

## Start with the pushbacks you actually get

You can't write a playbook for objections you haven't named. Most founders carry a rough idea of why people say no, and it rarely matches the words sitting in their LinkedIn inbox.

Open your sent folder and read the last 30 days of conversations. Copy every message where a prospect pushed back, went quiet, or asked a question you fumbled. Paste them into a doc without tidying them up. The exact wording matters, because the AI will pattern-match on it.

Sort them into four or five groups. Timing, budget, competition, authority and friction cover most of what happens on LinkedIn. A group with one example isn't a pattern yet, so leave it out until you have three.

Count how often each group shows up. The one that appears most often is where the playbook earns its keep, and it's the one worth testing first.

| | Objection group | What the prospect writes | What the AI should do |
| --- | --- | --- | --- |
| Timing | Not the right quarter for us | Ask what changes next quarter, then offer one small thing now |
| Budget | We don't have budget for this | Ask whether it's a freeze or a priority call, then offer a smaller start |
| Competition | We already use something for this | Ask what the current tool misses, without naming or knocking it |
| Authority | I'm not the one who decides | Ask who owns it and what that person cares about, then ask for an intro |
| Friction | Just send me some info | Ask which part they want to see so the reply isn't a generic deck |

## Write a response framework the AI can follow

Left alone, AI drifts to one of two failure modes: it apologizes for bothering the person, or it argues with them. A framework keeps it in the middle. Four moves, in order, every time.

Keep each move to a sentence. LinkedIn messages get read on a phone between other things, and a four-paragraph answer to 'we already have a tool' reads like a brochure.

1. **Name the objection in your own words.** Repeat the concern back in five words or fewer so the prospect knows you read it. 'So the timing is the problem' beats 'I completely understand your situation.'
2. **Ask one question that separates a blocker from a polite exit.** A company-wide budget freeze is a different problem from a project that lost its sponsor. One question tells you which one you're in.
3. **Answer with one specific thing.** Pick the single result that matches the objection. If they mention time, talk about the hours saved. If they mention a competitor, talk about the one job that tool doesn't do.
4. **Close with a small ask.** A yes or no question, a two-minute video, fifteen minutes on Thursday. Asking for a full demo on the first pushback ends the thread.

> Two attempts is the limit. If the prospect pushes back twice on the same point, the AI stops and the conversation goes to you or to a follow-up list.

## Build an objection library with real examples

A framework tells the AI the shape of a reply. Examples tell it what your replies sound like. Both go in the playbook, and the examples do most of the work.

Write each entry the way you'd explain it to a new hire on their first morning. What the prospect says, what you say back, and what you're trying to find out. Skip the marketing language.

Include the replies that worked and the ones that got you a polite 'thanks, not now'. Knowing what not to say is half the training.

- **them:** We already use another tool for our lead gen.
- **draft:** Makes sense. What made you pick them?
- **them:** Price, mostly. It does the basics.
- **draft:** The basics usually cover it until you're spending an hour a day sorting leads by hand. Is that happening on your side?
- **them:** A bit, yeah. But I don't have time for a call this week.
- **draft:** No call needed. I can send a two-minute clip of how a 40-person agency set the scoring up. Want it?

```
Objection: 'We already have a tool for this'
What it usually means: they aren't unhappy, they just haven't compared.
Reply shape: acknowledge the tool, ask what job it does badly, offer one example of that job.
Say: 'Fair enough. Most people who move to us had the same setup and the gap was X. Is X a problem for you?'
Don't say: 'Our platform is more powerful' or anything about features.
Stop if: they repeat the same answer twice.
```

## Test the AI on conversations you've already had

Before the AI writes to a stranger, feed it conversations you've already finished. Paste a real exchange into a test campaign and read what it writes back. You're checking two things: whether it picks the right entry, and whether it sounds like you.

The failures are usually obvious. It repeats the pitch after an objection, or it invents a feature you don't have. Each failure points at a line in the playbook that's too vague.

Run six or seven tests, one per objection group, and keep a note of what you changed after each. That note becomes the playbook's changelog, and it stops you from undoing a fix you made two weeks ago.

**Weak reply**

- Sorry to bother you again. I know you're busy, but I really think a call would help. Would you possibly have fifteen minutes?
- I hear you on price, but our platform is the most complete option on the market and includes more than anything else you're looking at.

**Strong reply**

- Fair. Is it the monthly cost or the setup time that's the issue?
- Understood. I'll send one page on how the scoring works and leave it there. If it's useful, we can talk next month.

## Decide what happens when the objection holds

Some pushbacks are real. The company is mid-migration, the budget is closed, the person isn't the buyer. The playbook needs an exit for those, or the AI keeps nudging people who already said no.

Write down what happens after the second attempt for each objection group. Most teams pick one of three options per group.

- Hand it to you: the conversation moves to your inbox and the AI stops writing.
- Park it: the lead goes to a group you revisit in a quarter, with a note about what the objection was.
- Close it: mark the lead as not a fit so no campaign touches it again.

## Keep the AI in your voice

The other decision is tone. If your buyers are technical, the AI should drop the pleasantries and answer in one line. If they're founders, a little warmth is fine. Put the instruction in the playbook in plain words.

Write the tone rule as a sentence the AI can act on: 'Short sentences. No exclamation marks. Never apologize for following up.' Vague instructions like 'be friendly' give you vague replies, and vague replies are what make a message read like software.

Add one line about what the AI must never claim. If a feature is on the roadmap, say so in the playbook, or the model will describe it as shipped and you'll be the one correcting it on a call.

## Doing this with Kinlead

Kinlead uses the playbook you've written when it writes messages and follow-ups for a campaign. You paste your objection entries into the optional sales playbook field, set the goal and tone, and the AI writes from there. Connection notes stay under 200 characters and open with why that person became a lead.

Each campaign chooses how replies are handled: hand the conversation to you at the first reply, let the AI reply until there's real interest, or let it reply until a meeting is booked on your scheduling link. Use hand-off for the objection groups you haven't tested yet.

Queued messages can be read and edited on the Actions page before they send, and any campaign can be paused. Replies from every campaign land in one inbox sorted by intent, with interested ones first. The [sales playbook guide](/ai-sales-playbook-for-linkedin-outreach) covers the rest of the setup.

## Related

- [AI sales playbook](https://getkinlead.com/ai-sales-playbook-for-linkedin-outreach.md)
- [AI sales script for LinkedIn messages](https://getkinlead.com/ai-sales-script-for-linkedin-messages.md)
- [LinkedIn outreach features](https://getkinlead.com/features/ai-linkedin-outreach.md)
- [Hand-off workflow](https://getkinlead.com/ai-sdr-hand-off-workflow.md)

## Common questions

**Should the AI try to close the sale or just book the meeting?**

Book the meeting. B2B deals close on trust built in a call, and a LinkedIn thread is a bad place to negotiate price. Give the AI a scheduling link and let it stop there.

**What if the AI gets an objection that isn't in the playbook?**

It falls back on the campaign goal and tone, which is why the fallback matters. Use hand-off mode for niches where the pushbacks are unusual, so a person answers the first reply.

**How often should I update the playbook?**

Every two weeks, or whenever a new objection shows up twice. Add the phrasing the prospect used and the reply you'd want to send, then test it on that exact conversation.

**Does a longer playbook work better?**

No. Ten sharp entries beat forty pages. The AI follows the closest match, and a crowded playbook makes the closest match harder to find.

[Create a Kinlead account](https://getkinlead.com/signup) or [book a call](https://getkinlead.com/demo) to talk through pricing and setup.
