How to Use AI for Lead Generation Without Sounding Like a Robot

You can tell within four words. "I hope this email finds you well." "I came across your profile and was impressed." Delete, delete.

The problem is not that the message was written by a machine. It is that it was written for nobody in particular and sent to everybody.

What to let AI do

Finding who to talk to. This is the part machines are genuinely good at and humans are genuinely bad at. Working through lists, checking who still exists, filtering out the ones who are wrong for you. Tedious, mechanical, and it takes a person days.

Working out what is true about them. Not writing the message. Gathering the facts. What they do, how big they are, whether they have just moved premises or hired or launched something. Give a machine that job and it will do it a hundred times without getting bored.

Following up. The reason most outreach fails is nobody follows up. Not because the follow up is hard, but because it is dull and there is always something more urgent. This is the single highest return thing to automate and almost nobody does it.

Sorting the replies. Interested, not now, wrong person, never contact again. Machines are fine at this, and it means the ones worth your time reach you quickly.

What to keep human

The actual sentence that says why you are writing. One line, specific to them, that could not be sent to anyone else. If that line is generic, the whole message dies regardless of how good the rest is.

Anything after they reply. The moment someone answers, you take over. Automated conversation is where trust goes to die. They wrote back to a person. Be one.

The judgement about who is worth pursuing. A machine will happily chase forty people who will never buy. You will look at the list and know.

The line, in one rule

Automate the work. Keep the words.

Everything that would be identical no matter who you were writing to, give to a machine. Everything that depends on this specific business, write yourself. That usually means a message that is ninety percent structure and ten percent you, and the ten percent is what makes it work.

Where people get it wrong

Volume as a strategy. Sending two thousand messages badly does not beat sending forty properly. It also gets your email domain flagged, which is expensive and slow to undo.

Letting it write the whole thing. You get grammatically perfect nothing. The kind of message that reads fine and lands with no impact whatsoever.

No human check before it sends. At minimum, read the first twenty. The mistakes AI makes are not spelling. They are confidently getting a fact wrong about a business, and that fact appearing in the first line.

Automating a broken offer. If people are not buying, the answer is not to tell more people faster. I say this constantly and it is the same reason I think you should fix the offer before paying anyone for marketing.

The honest version

I build this for a living, including my own lead generation software, so I have watched what works at close range.

What works is boring. A tight list of people who genuinely might need you. One true sentence about why you are writing. Two follow ups. A real person the moment they answer.

The businesses that struggle are the ones who bought a tool expecting it to replace having something worth saying. It does not. It just says the nothing faster.

If you are wondering where AI fits in your business more broadly, how a small business can actually use AI covers the wider picture.

If you want help setting this up so it sounds like you rather than like everyone else, get in touch. Or have a look at what I build.