McKinsey, Lightning AI, and Zscaler Redrew Three Services Org Charts. Your Vendor's Layoff Is Its Pricing Memo.
McKinsey is cutting up to 4,000 non-client roles over two years. Lightning AI named a former Heroku marketing chief as CMO, and Zscaler promoted its head of worldwide sales to CRO.
September 25, 2026
Three companies rewrote their org charts this week, and each rewrite points somewhere different.
McKinsey plans to cut non-client-facing staff by roughly 10%, or 3,000 to 4,000 positions, over 18 to 24 months. The cuts fall mostly on back-office and junior research roles, whose work generative AI can now finish in hours. Global managing partner Bob Sternfels has said the firm expects to need fewer staff outside client work. It will keep hiring consultants who work with clients directly. It's the largest reduction in the firm's roughly 100-year history. It follows Project Magnolia's 2,000-position trim in 2023 and about 200 tech-role cuts last November. Headcount is now down more than 10% from its 2023 peak of 45,100.
Lightning AI went the other way. On the same day McKinsey's plans surfaced, the AI-infrastructure vendor announced three senior hires: Betty Junod as CMO, Heather Paduck as CFO, and Jarlath Doherty as CPO. Junod previously led marketing at Heroku, VMware Tanzu, Docker, and solo.io. Paduck came from STACK Infrastructure, a hyperscale data-center operator. Doherty ran People at Squarespace and Collibra after global talent roles at LinkedIn, Yahoo, and Deloitte. The hires come nine months after Lightning AI's Voltage Park merger, which added 40,000 NVIDIA GPUs across nine US data centers. Together, they suggest Lightning AI is building for developers who ship products, with researchers who publish papers a lower priority.
Zscaler named Ross Tackett CRO effective October 1, promoting him from Head of Worldwide Sales. Tackett spent 10 years in sales leadership at ServiceNow and 16 at Dell. He succeeds Mike Rich, who is stepping down for personal reasons and will stay on as a strategic advisor through December 31. Zscaler ran no outside search and announced no re-plan or change to its sales motion.
What links three moves that look nothing alike
It's easy to file McKinsey's cuts under services and Lightning AI's hires under AI infrastructure. Zscaler's promotion would then go under security, and that would be the end of it. But all three companies are answering the same question: what should the org chart look like when the way work gets delivered changes? Each one is giving a different market a different answer.
McKinsey's answer is that the consulting pyramid has stopped paying off. The model relied on junior analysts to build the deck. If McKinsey's in-house tool Lilli can write it in an afternoon and 75% of employees use it regularly, the pyramid tilts. Non-client-facing work is shrinking because AI does it faster, and clients won't pay twice for the same billable hour.
Lightning AI's answer is that developers are the moat. The Voltage Park merger gave it compute, and then it hired a CMO from Heroku, VMware and Docker. That suggests it expects AI-infrastructure leaders to rise the way Heroku did in the 2010s, by winning developers. OpenAI's research-first rise in the 2020s is the other model. A company that hires a research-communications CMO after a compute merger is building for one audience. One that hires a developer-marketing CMO is building for another, so the profile of the person hired is the real signal.
Zscaler's answer is that its sales motion works and it would rather leave it alone. When a public security company has a soft pipeline early in its fiscal year, it usually brings in an outside CRO. The new leader resets the number, and the company absorbs the Q1 hit as a "new leader ramp." Promoting the head of worldwide sales instead tells Wall Street there's no re-plan and no compensation reset. More usefully, it tells the field that the playbook works and the job is to execute it.
Three signals for three different buyers
What you take from this depends on which side of the table you sit on.
If you buy consulting, read your services vendor's layoff announcement as its pricing memo. A firm cutting 10% of its non-client-facing staff has decided its delivery costs are going down. Whether that shows up in your next SOW as lower rates or as fatter margins for the vendor depends on whether you notice. At renewal, skip "are you using AI in delivery?" since everyone will say yes. Ask what the vendor's 2027 delivery model looks like, including how it will be staffed and priced. A vendor that can't answer has probably already built the savings into next year's proposal and hopes you won't ask.
If you sell to developers, or want to, watch which conference Lightning AI sponsors next. That's where the fight for the AI-infrastructure category plays out, and it's a more reliable signal than any positioning announcement. The last cycle followed a clear pattern. Docker had DockerCon and HashiCorp had HashiConf, while Heroku ran Waza and had a big presence at RailsConf. Junod ran marketing at two of those three companies. Whichever conference gets Lightning AI's top sponsorship in Q4 is the ground it plans to fight on. The vendors who thought they owned that audience will find they have company.
If you follow public security companies, read the promotion as a continuity signal. A public company can hardly send a stronger one short of filing an 8-K that raises guidance. Zscaler's next earnings report will read differently with Tackett in the job than it would have with an outside hire. On the next earnings call, watch whether analysts ask about sales-motion changes at all. If the continuity message landed, they won't, and a quiet call on GTM structure is a compliment nobody announces.
The counter-case worth taking seriously
The strongest objection is that moves like these happen every quarter, so calling this week's batch a pattern may just be pattern-matching. McKinsey has cut before, and Lightning AI hires senior people all the time. Zscaler had a CRO before Tackett.
That's a fair point, and it's why none of these moves is major news on its own. What makes the week worth noting is the timing. Three companies at different points in the GTM stack each made a bet on their delivery model on the same day. When that happens, it's usually because they've been looking at the same Q4 planning assumptions and reached their decisions at about the same time.
A related argument this week is that AI-native GTM vendors keep hiring the same kind of CRO. It's an operator who ran an enterprise-SaaS revenue org through the last shift to consumption pricing. Zscaler's promotion is the counter-move, a public security vendor with a working sales motion choosing continuity over that template. The template exists because most vendors need a proven playbook they don't have yet. Zscaler already has one, which is exactly why it promoted from within.
What this means for services buyers, AI-infra founders, and GTM leaders
If you buy services, raise this before your next SOW renewal or QBR. Ask the vendor to put in writing how its delivery model will change between 2025 and 2027. If it can't, price accordingly.
If you're an AI-infrastructure founder or marketer, decide by the end of Q4 which developer conference you'll own the sponsor slot at. Lightning AI made its bid this week, and that audience is now contested.
If you're a GTM leader at a public company planning a CRO transition, check whether your last two quarters justify the cost of an outside search. Zscaler's promotion is a reminder that "we're fine" is a legitimate message. The market notices when a company can credibly send it.
The org chart is where every operating theory eventually shows up, and three companies just showed theirs. My bet is that by year-end, two more services firms will announce their own version of the non-client-facing cut. One more AI-infrastructure vendor will hire a developer-first CMO in the Junod mold. Zscaler's promotion is the outlier in this batch, and it should be read that way.
If you buy consulting, read your services vendor's layoff announcement as its pricing memo.
Before your next SOW renewal or QBR, ask your services vendor to put in writing how its delivery model will change between 2025 and 2027. If it can't, price accordingly.
