In partnership with
One API. 1,000+ AI Models
Change a single line of code and unlock 1,000+ AI models, smart routing for cost and performance, and persistent agent memory.
GM to the Top 1% ☕
NEC just did something that should stop every seller cold. It cut a procurement negotiation from somewhere between three and forty eight hours down to under one minute, by putting a machine buyer on its side of the table.
AstraZeneca did the same to requisition checks. Two days became ninety seconds. Early agentic negotiations are landing twelve to eighteen percent better than the deals humans used to close, and agent software spend is racing toward 206 billion dollars this year.
Read that as the person carrying the bag. On your next big deal, the counterparty may not be a tired procurement manager. It may be software that has your last three quotes, every comparable in the market, and a mandate to grind.
💡 THE BUYER JUST STOPPED BEING HUMAN. YOUR PITCH DID NOT.
For your whole career, negotiation leaned on human friction. The buyer got tired. The buyer liked you. The buyer ran out of time before quarter end and took the deal on the table. Every closing move you own was built to work on a person.
The machine buyer erases all of it. It does not get tired. It does not feel the relationship. It does not fear the deadline, it uses the deadline against you. It models your floor from your own history and holds there while you sweat. Twelve to eighteen percent is not a rounding error. That is your margin, walking across the table into their savings line.
Here is the part that matters. A model can only optimize the variables it is handed. Give it a spreadsheet of price and seats and it will win, because price and seats are exactly what a machine grinds best. The only way to beat it is to put things on the table it cannot cleanly price. Risk you absorb. An outcome you guarantee. Certainty of implementation. The relationship that de-risks the entire purchase. Sell the un-modelable, or get modeled.
🔧 HOW TO SHOW UP WHEN THE BUYER IS A MACHINE
You cannot out-endure software. You can change what it is allowed to score.
1. Assume an agent is in the room: Walk in expecting your quote, your history, and every comp to already be modeled. Nothing you say is new information. Your value has to live above the data.
2. Put un-priceable things on the table: Outcome guarantees, risk transfer, speed to live, a named human on the hook. A model cannot cleanly discount what it cannot cleanly measure.
3. Refuse the price-only frame: The second the deal collapses to price times seats, you lose to the machine. Re-anchor on the business result every time it drifts back to a unit cost.
4. Sell to the human who owns the risk: Somebody still signs and still gets blamed if it fails. Find that person. The agent optimizes the number. The human owns the outcome. Sell to the outcome.
🎯 THIS WEEK'S HOMEWORK
Take your biggest open deal and list every variable currently on the table. Circle the ones a machine could price in ten seconds: unit cost, seat count, discount percent. Now write down three things you could add that a model cannot cleanly score. That list is your negotiation.
❓ QUESTION OF THE DAY
If a buying agent modeled your last deal, would it find anything on the table besides price?
Reply with the one thing you sell that a machine cannot put a number on.
🎁 REFER, AND ARM ANOTHER SELLER
You are at {{rp_num_referrals}} referrals. You are {{rp_num_referrals_until_next_milestone}} away from the next reward. Every seller you send here is one more person who reads the shift early instead of late. Your link: {{rp_referral_hub_url}}
See you tomorrow.
Edward
Founder, Morning Sales
P.S. Beating a buying agent runs on prep the machine assumes you will skip, knowing the buyer's business well enough to reframe the whole deal off price. I packaged the 500 prompts I use to do that fast, the deep account research, the value hypotheses, the reframe scripts. It is 27 dollars: https://store.edwardgorbis.com