Before Your CRO Reads the Coinbase Memo, Read This First
CEO Brian Armstrong sent an internal memo to Coinbase employees yesterday. Then he posted it publicly.
700 people gone.
Org structure flattened to five layers max. No pure managers. Every leader must also be an active individual contributor.
And a new operating unit called “one-person teams,” where a single employee runs AI agents to do the work that used to require a full pod.
Shopify did so too.
The company told employees: no new headcount unless you prove AI cannot do the job.
Klarna says its AI assistant now handles the equivalent of 700 support roles.
Duolingo went “AI-first” and told teams to rebuild workflows around AI before requesting a single hire.
Salesforce paused new engineering hires after AI tools boosted developer productivity by roughly 30%.
Greg Isenberg tracked it: 882 tech jobs disappearing per day right now.
That number will reach your CRO’s inbox this week. If it has not already.
What matters for CS leaders right now is that every company on that list described the same structural change.
Fewer layers.
Fewer managers.
Smaller teams doing more.
One person plus AI agents replacing what used to take five.
Your CRO did not read the Coinbase memo as a crypto story. They read it as an operating model.
And the first function they will pressure-test against that model is the one that sits closest to the customer and farthest from the revenue line on the org chart.
That is Customer Success.
The question your CRO is forming right now
It is a simple question.
Most CS leaders cannot answer it cleanly.
“How many people on your team do work that AI cannot do?”
Not “how many people are on your team” nor “what does your team do.”.
The question is narrower and harder: what specific work requires a human, and can you prove it?
The CS leaders who get stuck on this question stumble because they have never separated their team’s strategic output from their operational throughput in a format leadership can read.
Three places that gap shows up.
All three are visible to your CRO right now, whether you realize it or not.
1. Your team’s activity report looks automatable
Pull up your team’s last monthly report. Look at what fills the page.
Health score updates.
Renewal timeline tracking.
QBR deck assembly.
Onboarding check-in cadences.
Customer communication summaries.
Risk flag documentation.
Every one of those activities was important when a human had to do it.
Every one of those activities is now something an AI agent can execute faster, cheaper, and at higher volume.
When leadership looks at your team’s calendar and your team’s reporting, they see task execution. They see volume. They see activity.
They do not see:
The judgment calls your team makes between those tasks.
The political read your senior CSM made before escalating a renewal risk.
The exec relationship your team lead built over 18 months that nailed an expansion conversation no AI could have started.
That invisible work is real. It is also undocumented.
The Coinbase memo I mentioned above eliminated task execution roles.
The activities that fill your team’s reporting look identical to the work Armstrong just cut.
Your CRO will see the resemblance.
2. The “strategic advisor” narrative has no proof system
CS leaders describe their teams the same way in every leadership review.
“We are strategic advisors to our customers. We drive executive alignment. We surface risk early. We protect renewals through relationship depth.”
All of that can be true.
The problem is that when leadership asks for evidence, the answer is often some version of “it is hard to measure.”
That answer was acceptable two years ago. It was tolerated last year.
After the Coinbase memo and the Shopify mandate and the Klarna numbers, it sounds like a cost center that cannot justify its own headcount.
Jason Lemkin said it plainly this week: “If your team is truly AI-fluent, you should not have any managers who are not also ICs.”
He was responding to the Coinbase memo, but he was describing every function.
Including yours.
The CS teams that survive the next review cycle will have a proof system that connects their work to outcomes leadership already tracks.
Revenue protected.
Pipeline influenced.
Decisions changed at the customer’s executive level.
Churn risk retired with a documented sequence, not a relationship claim.
The ones that do not have that proof system will hear a version of the same question Armstrong’s team heard: “What if one person with AI agents could cover this?”
3. Your org chart looks like the one Armstrong just flattened
Director of CS.
Senior Manager.
Team Leads.
Senior CSMs.
CSMs.
Associate CSMs.
CS Ops.
CS Analysts.
Each layer made sense when it was created. Each role exists because someone needed it at the time. The problem is accumulation.
Leadership is now looking at every org chart through the Coinbase lens. Five layers max. No pure managers. Player-coaches only.
Count the layers between your most junior CSM and your VP.
If the answer is four or more, that structure is going to draw attention. The reason is leadership just read a memo from a $52B company that says it is unnecessary.
The harder question here is how many people in your org chart exist primarily to manage other people?
How many of them are also doing the customer-facing, revenue-impacting work themselves?
The “no pure managers” policy at Coinbase was about a belief that management without production is overhead.
Whether you agree with that belief is irrelevant. Your CRO might agree with it. And they just got a Fortune 500 case study that validates it.
The gap this exposes
The CS leaders who lose headcount in the next review cycle will share one pattern.
They described their team’s value in terms of what the team does. Activity. Coverage. Volume. Cadence.
The CS leaders who keep their teams will describe value in terms of what changes in the business when CS operates:
Renewal rates.
Expansion revenue.
Time-to-decision on at-risk accounts.
Executive alignment that shows up in contract velocity.
Those are two fundamentally different conversations.
Most CS leaders have only rehearsed the first one. And the first one is the conversation that sounds automatable.
If you have not yet mapped which track your individual contributors sit on, the diagnostic is here: The CS Career Split: Elevation Track or Automation Target?
The roles that survive this shift are already being priced. Here is what one of them looks like: Salesforce Just Posted a $350K CS Leadership Role. Here Are the 5 Skills It Demands.
What the next leadership review requires
Your CRO will sit down with a question they did not have six months ago.
The CS leaders who control that room will walk in prepared:
A scored audit that separates defensible work from automatable throughput.
A repositioning narrative in the language their CRO already thinks in.
An org structure argument that survives the “what if we cut this layer” test.
The system below builds all three.
It includes a workbook that scores your team’s activities, generates the risk heatmap, and produces a presentation-ready headcount justification.
Two ways to get it:
Upgrade to The CS Café Premium. The workbook is included with your paid subscription, along with every workbook in the archive and every new operating system I publish. Plus direct email review on your renewal plans, QBR narratives, and exec updates. Upgrade →
Or buy the workbook standalone: $49. Single download, no subscription. Get the workbook →
If you plan to use more than one system this year, the subscription pays for itself in the first month.
Looking for your next CS role? TopCSJobs.com has 460+ live CS roles updated daily, the only job board built exclusively for Customer Success professionals. 960+ CS candidates already on the platform. Browse the latest CS Jobs →