Deal Desk And Pricing: Using AI To Stop Discounting On Reflex

When a buyer says 'your price is too high,' the worst thing a rep can do is immediately discount. The second-worst is freeze. The right move is to slow down, understand what the buyer is actually telling you, and respond from a structure — not from anxiety. AI is unusually useful here because it never feels the pressure of an end-of-quarter close. A 10 percent discount on a $100k deal is $10k. If your gross margin is 70 percent, that $10k discount eats $10k of profit. To recover, you'd need $14k in new revenue at the same margin. Discounts are not a small concession — they're the most expensive thing a rep can give away.

• Up to 10% discount: rep approves
• 10-20%: manager approves with written justification
• 20-30%: VP of Sales approves with finance sign-off
• Above 30%: CFO and Deal Desk approve, requires multi-year or volume commit

Top reps almost never discount without getting something in return. Multi-year commit. Annual upfront payment. A reference. A case study. Public logo rights. Expanded scope at signing. Each of these has real value to your company and turns the discount into a structured trade instead of a giveaway. Before approving any discount, run the deal context through Claude or ChatGPT and ask: what's the most likely real reason behind the price objection (it's rarely price), what non-price concession could I trade instead, what's the smallest discount that solves the actual problem, and if I walk away, what's the realistic outcome?

A good rep can articulate, in two sentences, why their price is fair given the value delivered. They never discount without a corresponding ask. Their average selling price holds steady or rises over time, even as the company grows. AI doesn't replace pricing judgment — it gives reps a structured prompt to slow down and think before they reflexively cut their own margin.