The email arrived on a Tuesday with no subject line. It was addressed to six coaching vendors and contained two sentences⁰¹: the pilot had concluded, the decision had been made, a single replacement had been chosen. The replacement, attached to the email as a one-page architecture diagram, was not a vendor. It was an agent.

By the following Monday, four of the six had replied with variations of the same question: which agent? They were not asking about a product. They were asking about a category they did not yet know how to sell against.

§ 01 — THE SEAMHow procurement got there first.

For most of the last eighteen months, the story about agentic L&D has been told by L&D leaders. It has been a story about pedagogy, about human-agent collaboration, about whether a 1:1 coaching session can be replicated by a machine. It has been, in other words, a content story.

The procurement story is different. In March, the blended seat price for an agentic coaching license crossed $38 per employee per month across three major vendors — below the hourly rate of a mid-market executive coach for the first time⁰². For a 180,000-person industrial running six coaching pilots in parallel, the line-item math stopped being interesting and became obvious.

"We didn't run a bake-off. We ran a spreadsheet. The spreadsheet ran the bake-off." — Head of procurement, F-100 industrial · on background

This is the seam we did not see coming. Not the question of whether an agent can teach — which is a hard, slow, pedagogical question still being worked out across thousands of pilots — but the question of whether a CFO will wait for that question to be answered when the price curve is already below the line.

§ 02 — THE PATTERNThree peers, mid-RFP.

We've now confirmed — through primary interviews with people inside three peer companies⁰³ — that the same procurement pattern is underway at two other Fortune-100 industrials and one Fortune-50 financial services firm. In all three cases, the RFP has either been paused or converted into a sole-source agent-seat renewal⁰⁴.

What we cannot yet verify — and what we'll track across the next six briefs — is whether this compresses into a category reset by Q3, or whether one of the three stalls at procurement review, resets the pattern, and buys the human vendors another quarter. Both scenarios are live.

A contradiction we noted: one incumbent coaching vendor claims pipeline is 'strengthening' quarter-over-quarter⁰⁵. Our peer interviews do not support that claim. We are publishing the contradiction rather than resolving it.

§ 03 — WHAT TO WATCHThree signals for next week.

01 · The Workday tutoring release. Expected to ship in a gated rollout by 04.30. If Workday prices it as a bundled seat add-on rather than a standalone SKU, the category ceiling compresses again.

02 · The Q1 earnings commentary from three public vendors. Two of three reported by press time; the third reports 04.29. Watch for language shifts from "AI-assisted" to "agent-led."

03 · Our own Quarterly Report №14, shipping Q2. We'll resolve the category pricing curve with named companies, named buyers, and publishable seat-price benchmarks.