Essay
B2B selling is going through its biggest structural change in three decades: a mandate for selling teams to do things differently.
01 · The shift
A decade ago, the seller held the information: pricing, capabilities, customer stories. Buyers came to a meeting to learn, and the job was to inform. Now buyers arrive already knowing what we sell.
| The buyer, evolved | The seller, unchanged |
|---|---|
| Reads every review before the call | Still pitches what the buyer already knows |
| Watches every competitor demo on their own | Still opens with the same discovery |
| Talks to peers in private channels | Still scrambles for context mid-call |
| Asks AI for the comparison matrix | Still loses the week to busywork |
02 · The sales tax
For B2B organizations, reps are the engine that makes the quarter. Hitting the plan means hiring, ramping and replacing reps year over year. Does AI fix this, or just compound it?
$20M
of new ARR to find next year
34
AEs hired, ramped and kept in seat
30%
of ARR spent to get there
Illustrative model for a typical scaling B2B sales org.
If AI were really working for B2B selling, why is the sales tax still so high when CFOs report to Wall Street?
03 · The hidden cost of AI
Bolting on AI creates custom work to make everything fit. The real cost is the engineering to build it, the security to govern it, and the inference you pay to run it everywhere. It's metered, and it compounds.
Anthropic · Claude
One team's Claude bill jumped 3.5× overnight, the moment they crossed 150 seats. Usage never changed; the pricing tier did. Salesforce expects to spend about $300M on Anthropic tokens in 2026, most of it on AI-assisted software development.
Marc Benioff on the All-In podcast, May 2026, as reported by Techzine and TNW. Neither company has confirmed the figure in official disclosures.
Miro AI
AI is bolted on as credits, not seats: a diagram costs 5, an image 3, and none roll over. Run out and you buy an add-on. The AI bill sits on top of what you already pay per seat.
Miro pricing page and Help Center, checked October 2026. Plans include 25 AI credits per member per month on Starter and 50 on Business.
AnyTeam owns everything below the line.
04 · What it takes to win
The gap between winners and losers compounds every quarter. Winners become AI-first companies that reinvent how selling works. Losers do much of the same, and carry the tech debt for years.
| Area | Winners move first | Losers stand still |
|---|---|---|
| Memory | Build the company brain once, so memory compounds | Knowledge walks out the door when a rep leaves |
| Systems | One AI-native system across the stack | Bolt-ons stitched across 10–12 tools |
| Growth | Grow output per rep | Add headcount to hit the number |
| Selling time | Selling time climbs every quarter | Busywork stays and tech debt grows |
05 · The new scoreboard
The old scoreboard measured activity and headcount: calls made, reps hired, pipeline added. The new one measures whether the system makes each rep, and the whole org, structurally more productive.
Built on a new AI-native system: one company brain across the org.
A transparent value model, shown until named design-partner data is published.
Start building muscle memory now.