Don't use AI for your AI content RFP brief! Just kidding.

We've seen a bunch of AI video and photography RFPs in the last couple of months. Some of them good, some of them poor.

The short version: let more than one model review your brief, let a human read it last, and make sure the team behind it has actually produced content.

1. Let the models review each other

This sounds like common sense. Believe me, it is not always common practice. Write your draft with one model, then give it to another and ask one question: what would a supplier have to guess here? Let the tech bros (or the models within one tool) give each other feedback. They're good at spotting each other's blind spots.

2. Let a human read it last

No matter how many tokens you burn, bot language is not always humane language. Your future supplier is AI savvy, but might actually read your document with human eyes. So might your client, if you're on the other side of the table. Read it out loud once. Minor tweaks are welcome, and sometimes necessary.

3. Tech savvy is a start. Content savvy is a must.

Your brief tells a supplier whether it came from a content-savvy team or not. AI does a great job on most of the writing, but it helps if the people behind it have produced content for a while. They answer what a supplier would otherwise guess:

• What source material exists? Packshots, footage, technical drawings, brand assets.
• Who owns art direction, and who signs off?
• How many images or edits per product, per channel?
• Are real people, faces or voices involved, and who has cleared them?

Every blank becomes an assumption your supplier has to price in. Fewer assumptions, sharper quotes.

Second pair of eyes?

Book a free 10-minute call with me and Christopher Marks about your RFP. Two humans, zero bots. Promise ;)

Written by hand + keyboard. Tidied up for the website with a little help from the models it talks about.

First posted on LinkedIn, 21 October 2026.