Use AI to do the homework, not the mailshot

AI research before recruitment business development outreach

Some clients are already using AI to research the agencies they might call. Commercial director and AI adviser Nik Stapleton argues that agencies are often using the same tools the other way round, to send more, and sets out how to use them to know a client before the first approach.

Tony Restell wrote on this site last month that hiring managers are asking AI which recruitment firms they should consider, and that agencies are telling him enquiries are arriving because an AI engine recommended them. So some clients are already doing their homework on you with AI, before you've spoken.

Now look at how agencies are using the same tools. In Firefish's survey of 132 agency leaders, carried out at the end of last year, 43% named business development as a focus for AI, which it describes as tracking client signals and personalising outreach. The use I hear about most is the second half of that: more emails, written faster, sent to more people. So the client has used AI to understand you, and you've used it to send them a message that could have gone to anyone.

I'm not a recruiter, so I'll leave how to run a desk to the people who do it. I'm a commercial director, and a lot of my work now is AI: building it into how firms find and win clients, my own included. Before I contact anyone, it reads their accounts, their news and their adverts, and gives me the questions worth asking. I check every fact before I use any of it. That last part matters more than the tool.

More, faster, the same

When every agency's emails come out of the same tools, they read the same. The client can't tell yours from the next one, so they go back to the only thing they can compare, which is the fee.

The same survey found 54% of agency leaders putting more into the quality of their communication, in what Firefish calls an AI-saturated market. They're right to. What a client notices isn't whether AI wrote the email. It's whether it's relevant, and whether it reads like the last five they deleted.

Volume also hides the problem. A team that sends 400 emails and gets four replies can tell itself the number was the issue and send 800. AI makes that cost almost nothing, which is exactly why it's tempting.

What the homework looks like

None of this needs a new system. It needs a bit of time on each client you actually want, and one of the AI assistants anyone can buy.

Read their last twelve months before you write a word. Their filed accounts, their news page, their LinkedIn posts, the roles they've advertised. Ask the AI to summarise what's changed: a new site, a contract won, a leader arrived or left, a run of adverts for the same role. That change is your reason to get in touch, and it's the one thing a generic email can't have.

Ask it what you'd want to know in the first meeting. Give it what you've found and ask for the three questions that would tell you whether this client has a hiring problem worth solving, and how much it's costing them. You'll throw some away. The ones you keep are better than "what are your hiring plans this year?"

Check every fact yourself. AI is confident when it's wrong. A wrong fact in the first line of an approach is worse than no fact at all, because it proves you didn't look. If you can't find the source, don't use it.

Write one approach that only they could have received. One line on the change you found, one on why it made you think of them, one question. If you could swap the company name and send it to someone else, it isn't finished.

Keep what you found. Put it on the client's record, so whoever picks up the account next starts from what you learnt rather than from scratch.

A check you can run this week

Pull the last ten business development emails you or your team sent. For each one, ask a simple question: could it have gone to a different client with only the name changed?

If most of them could, AI has made your outreach faster. It hasn't made it better.

The uncomfortable part

Doing it this way means sending fewer. For a while, the activity numbers may look worse, and anyone measured on emails sent won't thank you for it. It also takes judgement, because AI will hand you plenty that's plausible and wrong.

But clients are starting to use AI to work out who you are. The agencies that win from here will be the ones that use it to work out who the client is, before they say a word.

Explore recruitment agencies and search firms suppliers

Related reading

Explore this subject in more depth.

Recruitment Spotlight: the people shaping recruitment. Read the interviews.The Onrec Awards: a night for the remarkable. March 2027. Explore the Awards.

Community poll

When should employers first share salary information with candidates?