Welcome back to AI Scale Tips. It's Thursday, which means we stop admiring the parts and wire them into a system that runs the standards you already set - even the ones you set by feel.

Today we open the hood on one solo consultant who took a 90-to-120-minute intake ritual and collapsed it into under 20. Not with a clever prompt pack. With three decisions about where her judgment lives.

In today's issue:

From: Three good prompts will save me time on intake.

To: Intake is a routing layer, and three prompts are where I decide where my judgment lives - not a shortcut, a container.

Here's the uncomfortable one. Your best intake prompt isn't the win you think it is.

Because a prompt is a tactic. It breaks when the model updates, and you're back in the room doing it by hand.

What survives is the architecture underneath: three explicit decisions about where judgment belongs. What the model handles. What the system flags for you to review. What never touches the model at all.

Three prompts isn't a shortcut. It's a decision about where your judgment lives and where the model's does.

Get the decision right and the prompts become interchangeable. Get it wrong and no prompt saves you.

A boutique consultant, two years in, 6 to 10 active clients. Her intake before the rebuild: a 20-minute prep ritual - LinkedIn, website, referral notes - then a 45-minute discovery call, then a hand-typed brief to herself before the proposal.

90 to 120 minutes per prospect. Before a single billable minute.

She rebuilt it as three prompts, each with a standing context layer above the model.

One: a qualification router. Intake form in, scored against her ideal-client criteria, out comes advance-or-decline with a rationale she can send verbatim.

Two: a pre-call brief. Form plus public info in, out comes 300 structured words - what they believe, what they've missed, the two questions that surface the real constraint.

Three: post-call synthesis. Raw notes in, structured proposal brief out.

Intake dropped to under 20 minutes. And the briefs got better.

Don't copy her prompts. Copy her three decisions. Do this today with your own intake:

1. Map the leak. Write down every step of your current intake and the minutes each costs you. You're auditing where your presence is load-bearing and where it's just habit.

2. Sort each step into three buckets. Model handles it. System flags it for your review. Never touches the model. That sort is the architecture. Everything else is just prompt wording.

3. Build the router first. One prompt that scores a short intake form against your ideal-client criteria and outputs advance-or-decline with a usable rationale. This is your highest-leverage decision - it protects your calendar before it fills.

4. Add a standing context layer. Above each prompt, paste your positioning, your criteria, your voice. When the model updates, this is what you keep. The prompt is disposable. The context is the asset.

5. Run it open-loop. AI drafts, you approve. Close the loop only once you trust it.

The consultant didn't get lucky with three prompts. She thought in layers first, then wired them. That's the whole game - and it's exactly what my free 3-part series walks you through, end to end, no product or audience required.

See the loop built, not just described: Watch the free 3-part series.

Every hour you spend on intake is an hour that doesn't compound. It's the fragility tax in disguise - a workflow built around your presence instead of a system that runs without you.

But here's what actually compounds: the standing context layer. The three prompts will get replaced. The next model, the next update, the next tool - the wording changes.

Your criteria, your positioning, your definition of a good-fit client? That's yours. Wire it above the model once, and every future prompt inherits it.

The prompts are the sprint. The judgment architecture is the asset. The system beats the sprint. Every time.

The goal was never to remove the human from intake. It was to remove the human from the parts of intake that don't require a human.

Notice the twist in the case study: her proposals didn't just get faster, they got better. Because the structure forced her to surface fit signals she used to miss when she was doing it by feel. The system didn't replace her judgment. It gave her judgment a repeatable container to run in.

That's the part worth chewing on. You're not automating away expertise. You're building a room where your expertise runs the same way every time, even on the days you're stretched thin.

Tomorrow, we pull the week up a level and I hand you one idea to anchor to.

One of the ways this newsletter makes money is through sponsored ads. Sponsors pay to put their offer in front of you, and every time you read or click one, it helps fund the free work that lands in your inbox every day.

Here's my promise: I'll only ever run a sponsor I genuinely believe is relevant and helpful to you. If it can't earn its place, it doesn't go in. With that said, here's today's sponsor.

AI agents are doing the work of full GTM teams. Pipeline, content, customer service. The startups using them aren't waiting for headcount approvals.

Get the free Practical Guide to Agentic GTM for Startups and see exactly where to start.

Unlock the Playbook

Before you go: Here are 2 ways I can help you scale smarter with AI

AiScaleTips is your founder clarity compass.
Most scale with chaos. You scale by design.

- Justin Glover