On April 14, 2026, ServiceNow eliminated its entire Quality Engineering function. Not some QEs. The function.
Several hundred engineers were given a binary choice: transition to a developer role or take severance. There was no reorganized QE team with a different name. No survivors in the original structure. The company decided it no longer employs people whose primary job is to test software the traditional way.
Then in June, ServiceNow cut hundreds more across solution consulting, product marketing, learning and development, and sales roles. The official company statement was unusually direct: "Our platform is generating real AI efficiencies inside our own business - we run the way we ask our customers to run."
That last sentence is the one every enterprise tech recruiter needs to sit with for a minute.
What "Deleting a Function" Actually Means
Standard layoffs cut 10%, 20%, maybe 30% of a team. Companies keep the function alive and hire back when conditions improve. Cycles repeat. Recruiters know this rhythm.
Deleting a function is different. It means the company has decided the category of work no longer has a home in its organizational structure. There is no hiring cycle that brings it back. When ServiceNow eliminated QE, it was not making a headcount decision - it was making a structural declaration: software quality assurance at enterprise scale is now an AI responsibility, not a human one.
ServiceNow CEO Bill McDermott followed the QE elimination with a public statement on April 22: the company will not backfill natural attrition through the end of 2026. The company is banking on AI productivity to hold headcount flat into 2027. Translation: the gaps ServiceNow creates through attrition will be filled by software, not by recruiting.
The Salesforce Precedent
ServiceNow is not the first enterprise software company to run this experiment on itself.
Salesforce deployed help.agentforce.com in early 2025. By September 2025, CEO Marc Benioff had confirmed what the numbers were already showing: Salesforce cut its customer service division from 9,000 to 5,000 people. Agentforce was handling more than half of all customer interactions. Support operating costs had dropped 17%.
Benioff's exact framing: "I've reduced it from 9,000 heads to about 5,000, because I need less heads."
Like ServiceNow, Salesforce positioned this as internal efficiency - proof that its own platform works. Like ServiceNow, the company simultaneously announced hiring growth in sales. The message from both companies is consistent: AI handles transactional support work, humans handle relationship revenue work.
Two of the largest enterprise software companies have now validated the same thesis on their own payrolls. That is not a coincidence. That is a market signal.
Which Functions Are in the Crosshairs
Based on what ServiceNow and Salesforce have announced, the functions at the greatest immediate risk inside enterprise software companies are not engineering core. They are the support layer that surrounds it.
Quality Engineering and QA. ServiceNow removed this as a job category entirely. AI-driven testing frameworks have advanced to the point where a dedicated human QE function no longer justifies its headcount at enterprise scale. This applies beyond ServiceNow - any company building software on modern AI-assisted platforms faces the same math.
Tier 1 and Tier 2 Customer Support. Salesforce eliminated 45% of its support division. AI agents handle case routing, FAQ resolution, and standard troubleshooting. Human escalation paths remain for complex issues, but the volume handling that justified large support teams has been automated.
Learning and Development - Content Production. The people who created training materials, documentation, and onboarding content are being replaced by AI systems trained on existing documentation. L&D roles focused on program design and organizational strategy are safer than those focused on content generation.
Solution Consulting - Rote Implementation. Consultants who primarily walk customers through standard product configurations face automation risk. Complex customization, multi-system architecture, and enterprise change management are safer. The rote part of the job is not.
Product Marketing - Content-Heavy Roles. Standard marketing copy, release notes, and product documentation are being automated. Strategic and positioning work requires more human judgment, but production-heavy roles are under pressure.
Which Roles Are Surviving - and Growing
Both companies are cutting and hiring simultaneously. The shift is not about shrinking overall spend. It is about reweighting it.
Sales. Salesforce grew sales headcount 20% year over year while freezing engineering. ServiceNow is shifting budget toward commercial functions. Revenue generation is the one role that has remained stubbornly human in every restructuring announced so far.
Enterprise Architects and Platform Strategists. Companies need people who design how AI agents interact with each other and with external systems. This requires organizational and systems judgment that current AI cannot provide. Demand here is growing.
AI Implementation Specialists. The role that barely existed three years ago. People who configure and orchestrate AI workflows inside enterprise platforms are now among the most requested profiles at companies deploying Agentforce, ServiceNow AI, and similar platforms.
Complex Customer Success. Account management that requires navigating internal politics, managing executive relationships, and coordinating across multiple stakeholders inside a client organization. AI cannot handle a meeting where the CMO and CTO disagree about the platform roadmap.
What Recruiters Placing Enterprise Tech Talent Should Do
The practical adjustments are not complicated, but they require acting before the compression is obvious in open req counts.
First, audit your most-placed job families against the list above. QE engineers, L&D content specialists, and Tier 1 support roles are already in structural compression at the industry level - not just at these two companies. The number of open reqs will continue declining across enterprise software, regardless of which platform the employer uses.
Second, retool how you assess solution consultants. The differentiating question is no longer "do they know the platform?" It is "do they know how to configure the AI layer on top of the platform?" A solution consultant who cannot speak to AI workflow design and agent orchestration is competing in a market that is shrinking fast.
Third, add one question to every discovery call with an enterprise tech client: "Which of your internal functions has your SaaS vendor automated away in the last 12 months?" The answer reveals exactly where that client's headcount pressure is coming from - and where it is going next.
Fourth, pay attention to where ServiceNow and Salesforce are actively posting. Both companies are building out sales, customer success, and AI infrastructure roles while cutting the functions listed above. If the vendors are growing a function internally, the talent demand for that function across their entire customer base is structurally strengthening. If the vendors are cutting it, their customers will follow.
The Vendor Proof Problem
When ServiceNow says "we run the way we ask our customers to run," the company is marketing its platform and restructuring its workforce in a single sentence. The subtext to every enterprise software buyer is explicit: our AI is good enough to eliminate these job categories inside YOUR organization too.
Every enterprise software company has a customer base receiving that message every time a vendor publishes internal efficiency gains, reduces its support headcount, or, as in ServiceNow's case, deletes an entire function off its org chart.
The typical lag between vendor proof and customer implementation is 12 to 18 months. That means the function eliminations happening inside ServiceNow and Salesforce today are previews of what arrives inside the companies that buy their platforms in late 2026 and into 2027.
Recruiters who understand that cycle can see the displacement before it shows up as cancelled reqs and frozen searches. The functions are identified. The timeline is visible. The only open question is which of your clients moves first.
If you place enterprise tech talent and want to see where active hiring demand is actually moving in your market, BlueLine's matching platform can surface the signal before it becomes obvious. Get started at bluelinesearch.ai/register.