The first proposal we wrote for this engagement was wrong. We rewrote it from scratch.

Last quarter we scoped a discovery engagement. Our first proposal was wrong. The client's inside champion told us. We threw it out and started over.

I'm writing this because most consulting firms do not talk about this part of the work. The part where the first interpretation of the engagement turns out to be the wrong shape. The part where you have to admit that the writing you just did is going in the trash. The part where you choose between shipping something flawed to hit the deadline or accepting that the correct response is to restart.

Here's what happened.

The firm had asked for a broad assessment of where AI-related opportunities existed inside their organization. Based on initial conversations with leadership, we formed a strong early hypothesis about which workstream mattered most. It was the thing we felt most confident building, based on the patterns in those first conversations. We wrote the first draft of the assessment around that hypothesis.

The inside champion at the firm read it and wrote back with a note that was, paraphrased: "This reads like a pitch for the one thing you're most interested in. That's not what we asked for."

He was right.

What the firm had actually asked for was a broad discovery assessment. They wanted to know where the highest-value opportunities were inside the firm, in priority order, with enough detail to make a real decision. Our early hypothesis was one candidate among many. They wanted us to evaluate it alongside everything else we'd found and present all of the options with equal weight, plus a recommendation on sequencing. Our first draft collapsed all of that into a narrative that led with our favorite opportunity because it was the one we felt most confident pitching.

The specific error is subtle and worth naming. We had done the discovery work. The pages reflected it accurately. The problem was the narrative shape. The draft told the story as "here's why you need the thing we picked, plus some other things you could also consider." What they needed was a story shaped as "here's what's happening in your firm, here are the highest-value opportunities, here's the case for each, here's our recommendation on sequencing." Same content, different spine.

Rewriting that kind of problem is not a light edit. You can't reshuffle sections. The frame at the top determines how every subsequent paragraph reads. We rolled the document back and rewrote from scratch.

Three days of work went in the trash. We billed zero hours for the rewrite because the rewrite was correcting our own error.

There's a version of this story where I would not have written it. The version where the narrative is "we work iteratively with clients to refine our deliverables based on their feedback." That version is true and also misleading. It implies the first draft was fine and we made it finer. The first draft was not fine. It was wrong. The correction was the work.

Most consulting firms don't tell this version of the story because it sounds like a mistake. It is a mistake. It's also unavoidable. When you're doing real discovery, your first interpretation of what the client needs is almost always wrong in a specific and important way. You don't know enough yet. You've built a mental model based on partial information and you're going to carry that model into the first draft because that's how writing works. The question isn't whether you'll get it wrong the first time. The question is whether you'll catch it before you ship.

In this case, the client caught it. That's fine. That's also how it should work. The inside champion at a client firm should be a second line of defense against your first interpretation. If the client can't or won't push back, the feedback loop breaks and you ship flawed work.

The healthier version of this process involves surfacing the interpretation early. Before we write the full draft, we should have sent a one-page framing document. "Here's what we heard. Here's the narrative shape we're going to use. Does this match what you asked for?" That would have caught the error two days earlier. I'm adding that step to the process going forward.

The broader point: a consulting engagement that never requires a framing reset is one of two things. Either the consultant is so experienced they nailed the interpretation in the first conversation, which is rare. Or the consultant is not actually doing new thinking for the client and is instead producing a standard product with the client's logo on top. The second one is more common than the first.

When you see a proposal come back exactly the shape you expected, ask yourself why. Maybe the consultant got it right immediately. Maybe the consultant brought a template. The difference between those two possibilities is whether the document teaches you something you didn't already know.

The rewritten version of our assessment taught the client something. The first version wouldn't have. That's the only test that matters.

Get your framing checked early. Be willing to throw away work when the framing is wrong. If your consultant is never willing to throw work away, they're probably not doing the work you think they're doing.