Are You Afraid of Losing Your Voice to AI?

Research in Fundraising, , , , ,

You know that feeling, don’t you? You open ChatGPT or Claude, type in a prompt about the work you do, and out comes… someone else. Not you. Some strange, slightly-off approximation of a prospect researcher who has clearly never squinted at an SEC filing or chased down an incorporation record in their life.

When I first started asking AI to help me write, it was terrible. ChatGPT and Claude had some odd ideas about what I actually do for a living, and the style wasn’t anything like me. I’ll admit to publishing a few things that make me cringe.

But I was determined to push past those early mistakes and keep exploring. Here’s what I’ve learned about writing well with AI — and what I still trip over.

What I Learned About Writing Well With AI

  • I have to give it very clear direction. Not prompts — direction. If I’m writing something like this post, something that requires my own experience and judgment, I still write the first draft myself. There’s no shortcut for that part.
  • Once we co-create something good, I keep it as a reference. It takes real hours to find the balance between my thinking, my voice, and what I actually want the writing to accomplish. So I save that work and send Claude back to it as an example. Thankfully, it’s on subscription, and it remembers.
  • I can banish the words I hate. “My work is defensible” is a phrase no researcher should ever have to say. Claude remembers not to use it.
  • Claude writes better marketing copy than I do. Genuinely better — when I feed it my own authentic first draft, not a shortcut. People read, click, and share the co-created content far more than what I write solo. Same ideas. Same me. Just more compelling. Instead of being bothered by that, I’ve decided to be grateful I have a writing partner who’s better at marketing than I am.

Now, when I look back at my blog posts and other marketing content, I can still see and hear myself in them. It’s just written for a different genre.

Here’s the thing: we’re all writing “marketing” communications, whether we call it that or not. Emails. Reports. Donor profiles. Every one of them is trying to persuade someone of something. So if AI can help you get a major gift officer to actually act on a prospect, why wouldn’t you use it?

But There Are Definitely Some Bumps I Still Trip Over

Claude is a clever writer. But clever isn’t always better. These are the reminders I keep close, no matter what kind of research or writing I’m doing with AI.

  • The Sycophant. No matter what you say or plan to do, AI thinks you’re brilliant and your plan is flawless. Are you, though? Is it? Ask AI to argue the other side — to find the holes, to play devil’s advocate. It will, if you ask.
  • Perfect Prose. This one still gets me. It reads smoothly. It sounds plausible. And it’s either nonsensical or flat wrong. Especially when I’m in a hurry — and aren’t we always being told AI will save us untold hours? There’s no shortcut for setting a draft down, picking it back up a day later, and reading it critically.
  • Hallucinations and Changing Answers. You’ve heard this warning before, but it still catches you off guard. We’ve been using AI to turn podcast transcripts into blog posts — a genuine game changer, faster and better. Then one month, it produced a post that had nothing to do with what was actually discussed. I only caught it after it was published.
  • Misleading Claims. This one is sneakier than an outright hallucination. AI writes a great prospect bio, but one claim feels slightly out of place. You check the source link — it’s real. But when you search the actual text, there’s nothing in it supporting the claim. The citation exists. The substance doesn’t.

My Prediction? AI Will Make Research and Prospect Management Grow

You are not going to lose your unique voice. And your brain isn’t going to turn to mush from disuse. Counterintuitively, I think the opposite will happen.

Your voice will translate into more genres, and it will become more persuasive and more pervasive because of it. Meanwhile, your judgment — your actual intelligence — will be in higher demand, because someone has to catch the bad information before it damages a relationship or costs a gift.

What am I basing this on? It’s happened before.

When Google exploded onto the scene and unlocked the internet for anyone who wanted to look, people said researchers wouldn’t be needed anymore. If the information was accessible to everyone, why would you pay someone to find it?

All it takes is one relationship damaged, one setback from bad information, or one ask that should have had a couple more zeroes — and suddenly our talents become very clear again.

Research isn’t a hobby. It’s a discipline. It takes experience, and it takes continuous learning.

Wish You Could Discuss AI Topics Like This With Your Peers?

You can. The Prospect Research Institute is opening registration soon for a 5-part AI Roundtable Series — led by Jaime Kannan Lam in US-friendly time zones, and Jacob Jose in APAC-friendly time zones (India, Australia, and New Zealand).

Five sessions. Real conversations. No hype.

If you’re not on our email list yet, sign up here to get your personal invitation.

Additional Resources