3:42am. The pricing email is half-drafted. The number that felt right at lunch on Tuesday — $2,500, dignified, fair to both sides — has, by 3am Friday, become I should have charged $1,200, they were right to push back, what was I thinking. The wave has moved, and the wave has taken the number with it.

By morning, you've sent the email at the wave-low number. By next Tuesday, looking at the work you actually have to deliver, you'll regret it. By month-end, you'll be tight on rent and tighter on energy. The number was never wrong. The wave was loud, and the body was alone with it at 3am, and the bot you were chatting with told you charge what feels good — which is the worst possible advice during a wave.

This pattern doesn't get fixed by charging more. It gets fixed by seeing the two columns separately while the wave is loud.

I wondered what would happen if a daily routine with an AI assistant — a workflow with AI — refused, on principle, to merge two questions that almost everyone collapses into one number under pressure. This is an inner workflow with AI: it doesn't manage the quote or the calendar. It holds the direction of your attention on a pricing day, when the body wants to look away.

Stage 6 — The energetic signal beneath the tight moment

Maybe the tightness isn't a confidence problem. Maybe it's the Scarcity set-point firing — the body's standing broadcast that switches on the instant a pricing decision comes near. On a clear Tuesday, the broadcast is quiet and the number arrives whole. By Friday at 3am, the wave is high, the broadcast is loud, and the same work re-prices itself downward without you noticing the swap.

This is a Phase 4 moment — the embodiment phase, where a new way of valuing your work has been installed but hasn't held under the visible-money threshold yet. The old pattern regresses right at the edge, the moment real money is on the table. And underneath it sits the mirror: the field returns what the body broadcasts, so clients meet the trying, not the price. The same protective patterns take turns running the drop — the Overgiver who shaves the number to stay liked, the Perfectionist who says it could be better before I charge this, the part that would rather under-charge than need anyone's help.

A workflow with AI can't quiet the wave. What it can do is keep the clear-day number pinned next to the wave number, so the two never merge into one. Tuesday's clarity stays legible while Friday is loud.

The Human Design chart, as the optional structural overlay underneath, names where this tends to live: an Open Heart center (the willpower Center on the BodyGraph (Human Design body chart) with no color — it over-promises to prove worth) running its Not-Self loop (the borrowed pattern an open Center picks up from others), and an Open Root center (the base pressure Center with no color — it takes in others' time-pressure unfiltered) that turns they'll find someone else if I don't reply tonight into personal urgency. Defined or open, the move is the same: hold Tuesday's clarity as a pinned reference, and don't let Friday over-write it.

Stage 7 — The question the AI asks first

Before any number, the partner asks two questions, separately. Never merged.

"What is this work worth?"

Worth to the client. Worth to the market. Worth in the founder's stated positioning. Three sub-answers, none of them a single number.

"What does this cost you to deliver?"

Hours, energy, recovery time. Opportunity (what won't fit in this month if this lands?). Material / subcontractor / tool cost. Three sub-answers, none of them a single number.

Then the third question:

"Are A and B the same number?"

Almost never. That's the work. The two-column split is the entire instrument.

Stage 8 — The workflow itself

A pricing assistant with two pinned references: a values inventory, and a log of the founder's last 6–10 pricing decisions with date + wave-state.

Every time a money decision is on the table — quote, rate, discount, invoice, late-payment chase — the workflow runs the two columns, separately, in plain view. Then a third move:

"What is the dignified floor, the stretch ask, and the walk-away line?"

Three numbers, not one. The founder reads them, picks one, logs the choice with a wave-state stamp.

What the bot does: refuse to merge worth and cost into one number. Refuse to recommend the number. Surface the pinned clear-day reference when the current wave is low. Log the final decision with date + wave-state.

What the founder still owns: the number. The bot cannot price. It can only hold the columns apart.

A concrete Friday morning, in four lines:

7:14am. Pricing decision: L's project, similar to last quarter's S work. Bot reads back the S decision: "$2,400, clear-day Tuesday, no regret at month-end." Bot pulls today's wave-state: low, but not 3am low. Bot lays out columns: – Worth: comparable peers $2,000–$3,200; client moves a $40K outcome; positioning supports $2,500. – Cost: ~30 hours; one weekend recovery; trades against the next discovery call. – Floor: $1,800. Stretch: $3,000. Walk-away: $1,400. Founder picks $2,400. Logs the wave-state. Sends at 8:02am. Reviewed Tuesday in clear; still feels right.

The bot did not optimise. It made the founder do the work of seeing two columns instead of one. The clarity came first. The number came after.

Stage 9 — The 7-day experiment that grows the capacity underneath

The workflow holds the columns apart. The experiment grows the capacity that makes the columns easier to see at all.

Choose-New-Words (7-day, money-tightness variant) — for seven days, remove one of should, behind, not enough, deserve, worth it. Notice what the body does when the word is unavailable. The hypothesis: tightness runs on a vocabulary that pre-loads scarcity into every financial thought. Removing one word doesn't fix the relationship to money; it surfaces how often the word was being used in place of clarity.

Daily signal: one note about a moment when the removed word would have appeared. Week-end review: which decision the founder thought was about money turned out to be about something else?

(Companion: Feeling Walk (7-day) — rebuild the valuing function narrowed to what does this generate? — especially in the Open Heart variant. Companion 2: the deferred-life audit — name what is being postponed until enough money arrives; surfaces whether the threshold can ever be met.)

Stage 0 — Return / Become

After two months of logged decisions, a pattern surfaces. The wave-state stamps are honest about what the body was doing on each pricing day. The Tuesday-clear decisions hold up at month-end. The Friday-low decisions don't. The data is data.

The bot doesn't charge more. It returns body-authority on each financial decision. The founder learns to wait through a wave by reading back the last clear-day reference, instead of pricing from inside the wave.

It's not pricing strategy. It's a refusal to merge two questions that don't belong merged.

The number stops being the pain. The number was always downstream. The work was learning to see the columns when the wave was loud — and the bot's whole job is keeping the columns apart, week after week, until the founder can do it without the bot.

What was tight starts to have edges. What was a wave becomes a wave again, with a beach to come back to.

This pattern fits an Open Heart center (no-color willpower center) and Open Root center (no-color base pressure center) — anywhere money-tightness has been running the nervous system. The bridge for non-pricing decisions is the reflection bot; the closing companion for the Sunday-evening Stripe-check loop is the shutdown companion. See your chart →