Personalised outreach AI that actually personalises. Not “hi {first_name}” nonsense.

Published March 23, 2026

This is part of our AI Lead Generation series.

Let’s talk about what personalised outreach AI actually means. Because right now, the bar is on the floor. Most “personalised” outreach is a template with merge fields. {first_name}, {company_name}, {industry}. Maybe a line about their latest LinkedIn post, scraped and paraphrased by a bot.

That’s not personalisation. That’s mail merge with lipstick.

Real personalisation means the recipient reads your message and thinks, “This person understands my situation.” It means the content of the message would only make sense for that specific person at that specific company at this specific time. Change any variable and the message falls apart.

That’s the standard. Anything less is just automated spam wearing a name tag.

The personalisation spectrum

There’s a spectrum that most people don’t see. On one end, you’ve got zero personalisation. Blast the same email to 10,000 people. On the other end, you’ve got a hand-crafted message that took 30 minutes to research and write.

Most personalised outreach AI sits about 20% along this spectrum. It adds the person’s name. It mentions their company. Maybe it references their job title. This is better than nothing but it doesn’t actually change the substance of the message.

What we build sits at about 85%. The message references specific, verifiable details about the prospect’s situation. It identifies a pain point that’s genuinely relevant to them right now. It connects that pain point to a specific outcome. The remaining 15% is the kind of nuance that only a human who’s had a conversation with the prospect could know.

85% is enough. It’s enough to get replies. It’s enough to book meetings. It’s enough that the prospect treats your message as a legitimate communication, not inbox pollution.

What “actually personalised” looks like

Let me show you the difference with real examples.

Fake personalisation: “Hi James, I see that TechCorp is growing. Companies like yours often struggle with scaling their operations. We help businesses with AI. Interested in chatting?”

This could be sent to literally anyone at any growing company. “Companies like yours” is the giveaway. It means “I know nothing about your company specifically.”

Real personalisation: “James, noticed TechCorp just opened a Manchester office and posted 12 roles in 3 weeks. Scaling that fast usually means your onboarding process is about to break if it hasn’t already. We built an automated onboarding system for a similar fintech that cut new hire ramp time by 60%. Want to see how it works?”

Every element is specific. The Manchester expansion (verifiable). The 12 job postings (verifiable). The onboarding pain point (logical inference from the data). The fintech comparison (relevant to their industry). The 60% metric (concrete and credible).

James can check whether you’re telling the truth. And that’s the point. Real personalisation is verifiable. Fake personalisation is vague enough that it can’t be checked.

How the AI gets to “real”

Personalised outreach AI that actually works needs three things.

Deep data collection

Not a LinkedIn profile scrape. A full research brief that includes company financials, hiring patterns, technology stack, market position, recent events, competitive positioning, and decision-maker background. This requires pulling from 5-10 data sources per prospect.

Inference logic

Raw data alone doesn’t create personalisation. The system needs to infer pain points from data signals. “12 new hires in 3 weeks” becomes “onboarding is probably overwhelmed.” “Switched from Marketo to HubSpot” becomes “they’re simplifying their tech stack, maybe unhappy with complex tools.” “Series B funding 4 months ago” becomes “they’re scaling and need systems that scale with them.”

These inferences are where AI adds genuine value. It connects data points to likely business challenges at a speed no human can match.

Voice matching

The output has to sound like you, not like an AI. This means training the writing model on your actual communication style. Your vocabulary. Your sentence structure. Your level of formality. If you’re casual and direct, the emails should be casual and direct. If your brand is more corporate, adjust accordingly.

The two-bullet framework

The highest-performing cold outreach format we’ve tested across hundreds of campaigns is dead simple.

Bullet one: a specific pain point the prospect likely has, based on a signal you detected. Bullet two: a concrete deliverable you can offer that addresses it.

Then a direct calendar link. No “would you be open to discussing?” Just a link to book time.

Here’s why this works with personalised outreach AI. The system has the research to make bullet one genuinely specific. And the inference logic to connect it to a relevant deliverable. The result is a 2-3 sentence email that communicates more understanding of the prospect’s situation than most salespeople demonstrate in a 30-minute call.

Reply rates on this format consistently hit 6-7%. In some verticals, we’ve seen 10%+. Compare that to the 1-2% that generic AI outreach produces.

Why generic “AI personalisation” is getting worse

Here’s something most people aren’t talking about. Generic AI personalisation is actively becoming less effective. Not just stagnant. Worse.

The reason is saturation. Two years ago, if you referenced someone’s LinkedIn post in a cold email, it felt personal. Nobody was doing it. Now everyone is doing it. Every AI outreach tool scrapes LinkedIn and adds a “saw your recent post about X” line.

Prospects have caught on. That LinkedIn reference line is now a signal that the email is automated, not that it’s personal. It triggers the spam instinct instead of the “this person did their homework” instinct.

This means the threshold for what counts as “personalised” keeps rising. Surface-level personalisation, the stuff most tools do, is now worse than no personalisation because it screams “bot.”

The only way to stay ahead is to go deeper. Reference data that most tools don’t access. Make inferences that require connecting multiple data points. Show understanding that couldn’t come from a 10-second LinkedIn scrape.

That’s what our personalised outreach AI systems are built for. Not to compete with other AI outreach. To compete with the best human sales development reps, the ones who spend 20 minutes researching each prospect.

Building versus buying

You can’t buy real personalisation off the shelf. The tools that claim to offer it are working with the same limited data sources as everyone else. LinkedIn, ZoomInfo, Apollo. The research is shallow and identical to what your competitors have access to.

Custom-built personalised outreach AI uses data sources specific to your market. According to McKinsey research, companies that invest in deep AI personalisation see 5-15% increases in revenue and 10-30% increases in marketing efficiency. It starts with AI prospecting that finds the right companies before your competitors do. Industry databases your competitors don’t know about. Public data that requires custom scraping. Signal combinations unique to your ICP.

The build takes longer. It costs more upfront. But the output is genuinely different. Your emails don’t look like everyone else’s emails. Your prospects respond because the message shows understanding that clearly required effort.

And Harvard Business Review research shows that businesses deploying sophisticated AI for sales and marketing automation are seeing significantly higher conversion rates than those relying on basic templated approaches. Once built, the system runs at scale. 20 minutes of research quality applied to 200 prospects a week. Without a single person manually researching anyone.

That’s not a minor efficiency gain. It changes what’s possible for a sales team of any size.

Frequently asked questions

What is personalised outreach AI?

Personalised outreach AI refers to using artificial intelligence to craft highly customised messages that are tailored to the specific situation of the recipient. This goes beyond simply inserting the recipient’s name or company - it involves referencing verifiable details about their current challenges and goals in a way that demonstrates a deep understanding of their unique circumstances.

How is personalised outreach AI different from generic email templates?

Generic email templates with merge fields like “{first_name}” and “{company_name}” do not constitute true personalisation. Real personalisation requires the message content to be specifically crafted for that individual recipient based on research about their current pain points and needs. The message should reference concrete, verifiable details that could only apply to that person.

What level of personalisation can I expect from your services?

Our personalised outreach AI solutions typically achieve around 85% personalisation. This means the message will reference specific, provable details about the recipient’s situation and connect that to a relevant outcome or solution. The remaining 15% is the type of nuance that can only come from a direct conversation with the prospect. This level of personalisation is enough to get high reply rates and book meetings.