Tell the Expert Before Their Work Becomes Your AI’s Benchmark

SF Scott Farrell • November 22, 2026 • scott@leverageai.com.au • LinkedIn

Using your expert's work as the AI benchmark without asking? That's taking, not tuning.

Here's the situation. Your best person hand-built a set of pages for a live sales lead. They're excellent. Meanwhile, you're building an AI engine to produce pages like that, and the engine needs a bar to aim at.

The pages are sitting right there. The expert is flat out. Why interrupt them?

Because the moment they find out their work became a training target nobody asked them about, the relationship changes. What was creative abrasion, two sides pushing each other toward better work, becomes extraction. Technically the signal still works. Socially, you've spent trust you'll need later.

The cost that's easier to miss is quality.

When you mine someone's work in silence, you decide which qualities to reward. Was it the structure? The tone? The way they framed the client's problem? You pick, then you treat your reading as their judgement. The engine learns your guess about their taste, not their taste.

The fix is small. Ask before you use the pages. Say what they'll be used to judge. If they agree, show them two versions side by side and ask: which would you keep, and why?

Fifteen minutes of that beats a month of guessing. You don't need them to write a rulebook. People can't explain their craft from scratch, but they recognise it instantly when it's in front of them. Their "why" becomes the standard.

The expert stops being a source you farm and becomes a co-author of the bar.

Who owns "good" in your AI work right now: the person who makes it, or the person who quietly decided what it meant?


Discover more from Leverage AI for your business

Subscribe to get the latest posts sent to your email.

[rpnnrecommend total="5" thumbnails="false" class="recommendedlist" ]

Leave a Reply

Your email address will not be published. Required fields are marked *

© 2026 Leverage AI, Scott Farrell. All rights reserved. This content is made available on a limited, revocable, read-only basis only. No licence or right is granted to copy, reproduce, republish, scrape, store, adapt, summarise, index, embed, or use this content to create derivative works, work product, deliverables, methodologies, training materials, prompts, templates, software, services, research, or commercial outputs, whether by humans or machines, without prior written permission. This restriction includes internal business use, client work, consulting, advisory, implementation, and any use in or for artificial intelligence, machine learning, data extraction, retrieval, evaluation, fine-tuning, or knowledge-base construction.