AI Is All Over Sales. The Results Aren't Moving.

Walk the floor of almost any sales org this quarter and you'll find AI everywhere. It writes the call summary. It updates the CRM. It researches the account, drafts the sequence, and sends more email than a rep ever could by hand. Then ask what it did for the number that pays for all of it, the share of sellers hitting their target. Almost nobody can point to a change.

BCG put that gap in plain terms this month. Ninety percent of the revenue leaders it surveyed report meaningful AI use in sales workflows, yet only 8% report higher quota attainment (BCG, The 2027 Revenue Agenda, October 2026). We've already written about what that means for the conversation with your board. This piece is about a different question: not whether AI paid off, but where it was pointed.

We automated the parts of the deal that were already easy

Look at where the tools went first. Data entry, CRM updates, account research, meeting notes. BCG's own read is that AI has gone furthest in tasks "with relatively clear inputs and outputs," and least into the customer-facing, judgment-heavy parts of the sale.

That made sense as a starting point. Those tasks are measurable, low-risk and easy to demo. But they were never where deals were lost. A deal rarely dies because the CRM was out of date. It dies in a conference room the seller never enters, when someone asks whether this is worth the money and nobody has a good answer.

So the industry got very good at speeding up the parts of the deal that weren't the bottleneck. The bottleneck stayed exactly where it was.

More volume is the most expensive way to stand still

The second big use of AI in sales has been reach: more sequences, more personalization tokens, more touches per account. It feels like leverage. In practice it's an arms race with no winner.

Practitioners who live in outbound are saying this out loud. One cold-email operator who says he's sent more than ten million cold emails put it bluntly on X this week: every founder selling into the same inbox now has the same models, so prospects get a stack of cold emails a day that read the same, and "the email costs nothing to write anymore" (@coldemailchris on X). His conclusion is that the work that still matters is choosing who to contact and what to offer them.

That's the right instinct, and it reaches past outbound. When AI makes a thing free to produce, the thing stops carrying information. Buyers learn to ignore it. A generic first line, a generic ROI slide, a generic "teams like yours saw" claim: all of it is now noise by default. The scarce asset isn't the message. It's specific, believable reasoning about this buyer's business.

The deal stalls at the business case

Here's what buyers are actually asking for. In the same BCG survey, 51% of sales leaders say buyers expect more customized proof of value, including evidence of ROI, before they purchase. Only 32% say AI is helping them speed up those proof-of-value moments.

That's the gap in one line. Buyers want a case built for their business. Sellers have AI for almost everything except building it.

And the business case is hard in a way that admin work isn't. It means pulling the buyer's own numbers into a model. It means stating assumptions a finance team can test. It means presenting projections, not promises, because the buyer's CFO knows the difference. Historically that work went to a value engineering team, if the company had one, and only for the biggest deals. Everyone else got a spreadsheet template and good intentions.

AI made the message free. That's exactly why the case built for one buyer is now the thing that stands out.

Point the AI at the decision, not the activity

If you lead sales or RevOps, the reframe is simple to say and hard to do: stop measuring AI by the activity it creates and start measuring it by the decisions it moves.

A few questions worth asking of every AI investment on your roadmap:

• Does it touch the moment the buyer decides? Notes and sequences help the seller. A defensible case helps the buyer's committee say yes.
• Does it produce something a finance team can check? Black-box numbers get discounted to zero. Visible assumptions get debated, and debate is progress.
• Does it reach every deal, or just the top ten? Value work that only the biggest opportunities get isn't a strategy. It's a staffing limit.
• Can your champion carry it into the room without you? Most of the real decision happens when the seller isn't there. See how to build a business case your champion can defend.

This is where AI value selling earns its name. The point isn't more AI in the sales process. It's AI aimed at the step where most deals actually stall.

What this looks like with Serre

Serre is value engineering for every seller. It helps reps build a business case for each deal, grounded in the buyer's own inputs, with assumptions shown in the open and outcomes framed as projections, not promises. You don't need a value engineering team to use it. Serre does the work an AI value engineer would do, on every opportunity instead of just the few that justify the headcount.

Turn every deal into a defensible case for change. Get started or contact sales.

Frequently asked questions

Why isn't AI improving sales results?

Most AI in sales has gone into admin work and outbound volume. Those gains are real but operational. BCG's October 2026 survey found 90% of revenue leaders report meaningful AI use, but only 8% report higher quota attainment. The commercial lift has to come from the customer-facing parts of the sale, where buyers decide.

What should sales teams use AI for instead?

Point it at the moments that shape the buying decision, especially the business case. Buyers increasingly expect proof of value specific to their business, and a clear, testable case is what helps a buying committee move forward.

Does more AI-generated outbound still work?

Volume alone is getting less effective because every seller has the same tools and buyers see the same messages. Practitioners report that relevance, meaning who you contact and what you offer, now matters more than how many emails go out.

Do I need a value engineering team to build business cases?

No. Value selling software like Serre lets every seller build a business case from the buyer's inputs, with visible assumptions and outcomes framed as projections, not promises, without a dedicated value engineer.