The best AI cold email personalisation tool depends on one fork: do you want a writing layer that researches each prospect and drafts a personal first line, or a full sending platform with an AI writer built in? If you already send from a tool you like, add a writing layer; if you're starting from scratch, one platform keeps research, copy and sending together.
A merge tag that drops in someone's first name and company stopped fooling anyone years ago. What buyers reply to is a line that could only have been written to them: the hire they just made, the page they launched, the problem their job ad hints at. AI tools now draft that line for you, but they come in two very different shapes, and picking the wrong shape is the expensive mistake here, far more than picking the wrong brand.
The fork that decides everything
On one side sit writing layers. They take a list of prospects, go and read about each one, and hand back a personalised opener or a whole draft. They don't send anything. You paste or sync the copy into whatever you already mail from. That's their strength: you keep your warmed-up mailboxes, your sequences and your reporting, and you only swap the bit that writes.
On the other side sit sending platforms that have grown an AI writer. Here the research, the copy, the mailboxes and the follow-ups all live under one login. You lose some choice on the writing side, since you get the writer that platform ships, but you lose a whole step of exporting and importing, and nothing goes missing between tools.
There's a third thing in this space worth naming so you don't confuse it with either: data and workflow platforms. They pull live signals and enrichment for each lead and can feed a prompt that writes the line, but they're a build-it-yourself kit more than a finished writer. Clay sits here, pricing by actions per month on its own pricing page.
A quick way to know which side you're on: count the tools you'd have to log into to send tomorrow's campaign. If the answer is already one and you're happy with it, you want a layer. If the answer is three and none of them talk to each other, you want a platform.
Writers and research tools that draft the personal line
SmartWriter
A hosted AI writer that turns prospect data into personalised cold emails in bulk, so you get volume without writing each opener by hand.
Twain
Writes researched, one-to-one outreach for you, then the reply-focused follow-ups that come after it.
Regie.ai
An AI prospecting platform whose research and messaging step studies the account and the buyer, then drafts in your own voice.
Clay
If your edge is data, this enrichment and workflow platform turns research and live intent signals into a personalised line for every lead.
Sending platforms with personalisation built in
Mailshake
Cold email and sales engagement software with an AI email writer, suited to sellers who want copy generated and sent from one place.
Salesforge
Running email and LinkedIn side by side? This outbound platform bundles built-in AI personalisation and an AI SDR with unlimited mailboxes and LinkedIn senders.
lemlist
A multichannel outbound platform where AI agents research leads and tailor the copy for each one around the trigger that made them worth contacting.
Woodpecker
Built for one-to-one campaigns across cold email and LinkedIn.
Instantly
A sales engagement platform with its own AI that writes and personalises outreach at scale, for teams who want one tool to write, personalise and send their campaigns.
Mailmeteor
For a personalised mail merge without leaving Gmail, Outlook, Excel or Google Sheets, this add-on sends from where you already work.
What the tool has to know before it writes
Every AI writer is only as personal as the facts it's fed. Give it a name, a job title and a company domain and you'll get flattery about the company's mission. Give it a dated reason to talk, such as a funding round, a new hire, a product launch or a page they've just changed, and the line suddenly earns its place.
When you compare tools, ask where the research comes from. Some tools read the prospect's site and public profile for you. Some pull intent or trigger signals and build the email around them. Regie.ai, for instance, says it researches each account and buyer and drafts using dated why-now signals. Others expect you to bring that data in a column and simply rewrite it nicely.
Want AI doing the heavy lifting in your marketing?
I build the systems that handle the boring 80 percent, so you get your week back. Done properly, with the human kept in.
If your list already has a strong signal column, a plain writer will do. If it doesn't, pay for the research step, because a beautiful sentence built on nothing reads as obvious filler the moment a buyer sees it.
Run this test in an hour
Demos always look brilliant because the vendor picks the prospects. Use your own list instead, and run two tools head to head: one writing layer and one sending platform's built-in writer.
- Pick twenty prospects you'd email this month anyway: ten you know well and ten you've never researched.
- Export the same columns for all twenty, including name, role, company website and, if you have it, one recent thing that happened at the company.
- Give both tools the same offer and the same instruction for the opening line, and generate a first line for every prospect.
- Paste the forty lines into one document with the company names hidden, then ask of each one: could this only have been written to this person?
- Check every fact a line states against the prospect's own website or profile, and mark any line you can't confirm as a fail, however good it sounds.
- Score each tool on lines that are both specific and true, then time how long it takes you to fix the rest to a standard you'd sign.
- Pick the tool that leaves you the least editing on the ten strangers, because that's the half of the list it'll write for in real life.
Expect the well-known ten to flatter every tool, since you'd have written a good line for them yourself. And a confident wrong fact is worse than a dull true one, because a prospect who spots it won't reply to anything else you send.
Mistakes that make AI personalisation read as fake
The personal touch lands on the wrong part of the email. A tailored first line followed by the same five-paragraph pitch everyone else gets still reads like a template. The personal bit should set up why you're writing, and the ask should follow from it.
A writer left to guess will invent something. If there's no signal for a prospect, tell the tool to fall back to a plain, short opener rather than invent a compliment about their culture. A lot of the cringe in AI cold email comes from a model filling a gap.
Sending before reading. At volume it's tempting to approve a batch unseen. Read a sample from every run, and read more of the strangers than the familiar names. If you're still learning to write these emails with AI help, this guide to writing cold emails with AI covers prompting and editing in more depth.
No personal line fixes deliverability. It won't rescue a cold domain or a burned list. If your mail isn't landing, sort the sending setup first; the cold email software comparison covers that side.
Do you make AI cold email personalisation tools? Tool makers can ask to be considered for a featured place on this list.
What you'll pay for, in words
Starting plans, where a pricing page lists one, run from about six dollars a user a month for a mail merge add-on to roughly a hundred and seventy dollars a month for a data platform. Woodpecker and Instantly showed no starting price on the pages checked for this list. Writing layers and data platforms tend to bill by usage: you buy a bundle of credits or actions each month, and every researched prospect or generated email spends some of them. That's fair if you send in bursts, and painful if a messy list burns credits on prospects you'd never have mailed.
Sending platforms usually charge a monthly plan, sometimes cheaper if you pay for the year, with the AI writer included and limits set by contacts, mailboxes or channels rather than by each line it writes. Mail merge add-ons are simpler still and charge per user.
So work out your monthly volume before you compare plans. A credit plan that looks cheap for a small test can cost more than a full platform once every prospect needs research, while a platform plan is money wasted if you only write a few dozen high-value emails a month. Plans and inclusions change, so check each vendor's pricing page on the day you buy.
Which type to pick
If you already have a sending tool with warmed mailboxes and you're happy with your sequences, add a writing layer and keep everything else. Your risk is low, and you'll see from the next few campaigns whether replies move.
If you're starting outbound from nothing, or you run email and LinkedIn together, pick a sending platform with AI built in. One login, one place to read replies, no copy going missing between tools. Before you commit to one, compare the other multichannel platforms.
If your advantage is better data than your competitors, a data and workflow platform is the one to build on, as long as someone on the team enjoys building. And if you send a small number of careful emails from your own inbox, a mail merge add-on with your own written opener will beat every AI here on tone.
Want your tool on this list? Featured spots and reviews here are paid partnerships. Here's how to partner with me.
