The Compliment Is Not the Research

The emails had been researched.

They knew the product names. They mentioned recent launches. They addressed the right sectors. Each one contained a tidy little sentence proving that somebody had visited the website rather than firing a generic message into Bristol with a blindfold on.

They were also dreadful.

Change the company name and a few nouns, and the same email survived perfectly:

I came across your innovative work. My background sits between several useful things. I build AI-assisted workflows. If this sounds even loosely useful, perhaps we could have a quick chat.

This is a peculiar kind of failure because all the facts can be correct. The problem is not hallucination. The problem is that the research has been used as decoration.

A company compliment is not a reason to contact the company.

“Your work is impressive” tells the recipient nothing except that their homepage loaded successfully. A paragraph listing five capabilities tells them even less. Market intelligence, content, CRM, process, AI, research, sales support: once every skill enters the email, no particular person remains inside it.

The message starts sounding like a small agency with no clients.

The fix was not warmer adjectives. It was to stop writing emails.

We went back to the companies and asked different questions. What is actually happening there now? Has a product moved into its first customer deployment? Is a new partnership changing how the company reaches buyers? Has the team quietly placed a “please standby” notice beside a forthcoming role? Is there a named person close enough to the work that the conversation belongs with them?

Most importantly: is there a reason to write at all?

That last question culled more than half the draft pile.

Some companies were interesting but quiet. Some were plausible matches with no current opening in the story. One had a live vacancy, which made speculative outreach less sensible than using the recruitment route. Another had a perfect-looking executive contact who turned out to be the wrong person; the useful conversation belonged two levels closer to the actual work.

Research became useful when it was allowed to say no.

The surviving notes got shorter. They stopped explaining the companies back to themselves. They stopped pretending to know internal problems from outside. They stopped carrying a miniature CV in the second paragraph.

One began with a line on a careers page and a plain question about the work forming around it. The sender said directly that he was not looking for the advertised sales function, then named the adjacent marketing, customer-research and operational work he actually wanted. No “solutions-focused professional.” No apology for not attaching a CV. No offer to unlock synergies.

Just a person who had noticed something real and knew why he was writing.

That distinction matters more now because producing plausible outreach is nearly free. A model can glaze thirty companies before lunch. It can make each message factually different while leaving them spiritually identical. Personalisation at that speed often means changing the wallpaper in the same empty room.

Good outreach is not a mail-merge with better source material.

It is selection.

This company, now. This person, for this reason. This part of my experience, not all of it. One honest question or possibility. Enough context to make the message legible, then stop.

Sometimes the research earns an email.

Sometimes it earns silence.