Why Customer Support Hiring Collapsed 65% in 2 Years [2025]

The Biggest Shift in SaaS Hiring Nobody's Talking About Clearly Enough

Something remarkable happened in the past two years. Companies stopped hiring customer support representatives at the scale they used to, and the numbers tell a story that most executives still haven't fully processed.

The data is stark. In Q4 2023, customer support roles represented 8.30% of all new hires across a dataset of 386,500 positions. By Q3 2025, that number had collapsed to just 2.88%. That's a 65% decline in two years. And here's what makes it more striking: almost half that drop happened in the most recent three quarters alone. The acceleration is real.

But here's what confuses people. Most companies aren't actually reducing their support capacity. In fact, many are expanding it. They're handling more customer questions, more tickets, more edge cases than ever before. The difference is that machines are handling the work instead of humans.

This isn't theoretical. This is happening right now at companies you know. At SaaStr, the organization that published this data, they went from 20+ support employees down to just 3 humans, replaced by AI agents handling the volume. The kicker? Their revenue growth went from negative 19% year-over-year to positive 47% during that transition. They're doing more with less because the less is now artificial intelligence.

What makes this moment important isn't just that customer support is changing. It's that support is the canary in the coal mine. It's the first major function where AI actually works well enough to replace headcount at scale. And that means engineering, sales development, and marketing are coming next. The hiring landscape of 2027 won't look like 2025. It'll look fundamentally different.

Let's dig into what's really happening, why it matters, and what you need to do about it right now.

TL; DR

• The 65% decline is real: Customer support hiring dropped from 8.3% to 2.88% of new hires in 24 months, with acceleration intensifying in the last three quarters
• Companies aren't reducing support quality: They're restructuring how support works, with AI handling 40-60% of volume and humans focusing on complex escalations and strategic accounts
• The role of humans is changing: Support roles are shifting from
  $50K-$
  65K generalists to
  $90K-$
  120K specialists focused on strategic relationships and technical problem-solving
• Support and success are merging: When AI handles routine tickets, remaining humans increasingly do customer success work like proactive outreach and expansion
• This is just the beginning: Support is the leading indicator. Sales development, engineering support, and parts of marketing will follow similar hiring curves over the next 2-3 years

Customer support roles as a percentage of new hires have declined from 8.3% in Q4 2023 to 2.88% in Q3 2025, indicating a significant shift towards AI-driven support solutions.

Why Customer Support Was Always Going to Be First

If you think about which business functions are most vulnerable to AI disruption, customer support probably doesn't immediately jump out. It requires empathy, nuance, product knowledge, and the ability to handle unexpected situations. Good support is genuinely difficult work.

But that's exactly why support is the first place where AI actually works.

Here's the paradox: support isn't vulnerable because it's simple. It's vulnerable because it's predictable. When customers reach out, they're mostly asking a finite set of questions. Password resets. Integration help. Billing questions. Onboarding guides. Feature requests. Basic troubleshooting. These aren't novel problems that appear once in a million conversations. They're the backbone of every support volume.

The math is actually pretty straightforward. Studies from companies implementing AI support systems consistently show that artificial intelligence handles between 40% and 60% of inbound ticket volume without degrading customer satisfaction. At some companies, it's higher. Gorgias, which built an AI-native support platform, is growing their AI agent business at 100% annually at roughly $100 million ARR. That's not a side project. That's become the main product.

What makes this work at scale is that customers don't actually want to talk to a human for simple problems. They want answers. Fast. If an AI can reset a password in 30 seconds versus waiting five minutes for a human agent to pick up the ticket, customers prefer the AI. It's not about replacing relationships. It's about eliminating friction from routine interactions.

The second reason support is first is that the ROI is immediate and measurable. When you implement an AI support system, you instantly see which tickets it can handle and which it escalates to humans. You get clear data on deflection rates, resolution times, and customer satisfaction. Compare that to, say, sales development, where determining whether an AI-generated lead is valuable takes months to measure. Support gives you answers fast.

And third, there's a maturity curve. The AI tools for support are actually really good now. They're not experimental. They're production systems handling billions of conversations globally. Intercom's Fin is handling real tickets at real companies. Zendesk's automation is embedded in their platform. The technology isn't a promise. It's already here.

What surprises most people is how well these systems work without constant human intervention. They don't need supervision for every decision. Modern AI support agents can handle complex flows, understand context across conversation threads, and know when something truly needs escalation to a human. They're not chatbots. They're something genuinely different.

The Numbers Behind the 65% Collapse

Let's look at what the hiring data actually tells us, because the numbers are more revealing than the headline.

When you see a 65% decline in hiring for a category, the normal assumption is that the category is dying. Companies are doing less of it. Revenue is declining. It's a contraction.

But that's not what's happening in support. The total spend on customer support isn't going down. In many cases, it's going up, just in different directions. What's shifting is the composition. The mix of how that money is allocated.

Traditional support software vendors are experiencing this directly. Intercom, one of the largest platforms, made a deliberate strategic choice to go all-in on Fin, their AI agent product. Their non-AI support business isn't growing anymore. It's flat or slightly declining. But the AI agent business is growing so fast that the overall revenue is expanding dramatically. Zendesk is experiencing the exact same pattern. The traditional ticketing and helpdesk business has matured. The growth now is entirely in AI automation.

In other markets, this is even more dramatic. Gorgias, which focuses on e-commerce support, has their AI agent business approaching $100 million ARR and growing at 100% year-over-year. Meanwhile, traditional support work at e-commerce companies is becoming a smaller percentage of the organization, handled more often by AI systems than by humans.

The hiring decline reflects this shift. Companies aren't eliminating support budgets. They're redeploying them. Instead of hiring 10 new support generalists to handle volume growth, they're implementing an AI agent to handle 60% of that volume and hiring 2 really senior support specialists to handle escalations and edge cases. The total headcount goes down. The total capability might actually go up.

What's important to understand is that this isn't a temporary efficiency play. This is structural. These aren't companies trying an AI tool for three months and going back to humans. This is companies fundamentally changing the architecture of how support works. When that happens, hiring patterns change permanently.

The shift towards AI in customer support is evident with 65% of spending now directed towards AI solutions, reflecting a strategic redeployment rather than budget cuts. (Estimated data)

How The Best Companies Are Actually Restructuring Support

If you talk to companies that are scaling from

First, they're shrinking the generalist support rep role. That traditional tier-one support person who handles a wide range of questions is becoming smaller in number. The companies that are scaling effectively have realized that when AI handles password resets and standard troubleshooting, you don't need 15 people to cover that work. You need the AI, plus maybe one person to handle exceptions and improve the system.

Second, they're expanding the specialist support role. The humans who remain are different. They're more technical. They have deeper product knowledge. They make more money, usually somewhere in the

Third, the role of support is shifting toward customer success. When AI handles 50% or more of routine tickets, the remaining human work looks less like "support" and more like "success." It becomes proactive outreach. Onboarding help for new customers. Expansion conversations. Understanding why a customer is stuck and helping them get unstuck. The functional lines between "customer support" and "customer success" are blurring dramatically.

Some companies are making this explicit in how they organize. They're combining the teams. Other companies are keeping them separate but having the support team do more success work. Either way, the traditional model where support is purely reactive (customer has problem, support solves it) is becoming less accurate. The model now is more like (AI handles routine issues, humans drive proactive relationship value).

Fourth, AI is becoming tier-one support officially. At companies like Klarna, at Intercom's customer base, and at dozens of SaaStr community members, the structure has flipped. The AI system is the first responder. It handles the ticket, and if it can't, it escalates to a human. This is different from how support worked before, where humans were tier-one and you maybe had some basic automation. Now the hierarchy is inverted.

This structural change matters because it changes the hiring profile. You're not hiring 10 people with customer support experience anymore. You're hiring 2 people with technical depth and customer empathy. You're hiring someone who understands how to manage and improve AI systems. You might be hiring a data analyst to understand support metrics. The skill mix is completely different.

The AI Actually Works. That's The Real Story Here.

There's a tendency when discussing AI disruption to talk about the technology as if it's theoretical. As if we're debating whether AI will eventually be good enough to replace humans in support. That misses the mark entirely.

AI support is working right now. Not in pilot projects. Not at a few companies. At scale. Globally.

When Intercom launched Fin, their AI support agent, they weren't launching a beta feature. They were launching a product that companies immediately started using to handle real customer conversations. The deflection rates (the percentage of tickets AI handles without escalation) are in the 40-60% range at well-implemented deployments. That's not impressive for the 10% of easy questions. That's impressive because it includes moderately complex questions that humans used to think required a person.

The customer satisfaction data is similarly surprising. Customers don't report worse experiences when they interact with an AI support agent, provided the agent is actually helpful. If anything, they prefer it in many cases because the response is immediate. No queue. No waiting for a human to become available. Just instant answers.

There are categories where AI still struggles. Highly technical edge cases that require deep debugging. Situations where a customer is upset and needs genuine emotional intelligence. Problems that require understanding the unique context of that specific account. But here's the thing: those problems don't represent 60% of support volume. They represent maybe 20%. AI handles the other 80%, and does it competently.

The economics of this are brutal for the traditional support model. If you're a company thinking about whether to hire five new support people or implement an AI system, the AI system usually wins on both cost and speed. It costs less. It's faster to deploy. It doesn't need training. It doesn't take vacation. It doesn't have bad days.

Some companies worry this will damage customer relationships. The opposite is usually true. When customers can get instant answers to standard questions, they're happier. They reach the human support person faster when they need genuine help. The humans can focus on customers with complex problems or high account value. Everyone's experience improves.

That's why the hiring collapse is happening and accelerating. It's not hype. It's not fear. It's companies making a rational economic decision because the AI actually works.

Which Functions Follow Support? The 2-3 Year Horizon

If customer support is the leading indicator, what comes next? Which functions will experience similar hiring curves over the next few years?

The answer depends on applying the same logic that made support vulnerable. Which functions have high volume, predictable patterns, and work that doesn't require deep strategic thinking?

Sales development is next. Sales development representatives spend a large portion of their time doing research, sending outreach messages, and qualifying leads. These are tasks with clear patterns. Does the company match the ICP? Is the timing right? Is there likely to be budget? AI can handle this. In fact, it already is at forward-thinking companies. Platforms that help with SDR work are growing. The hiring collapse in sales development will likely follow support with a 12-18 month lag.

Technical support and documentation functions are vulnerable too. When you can use AI to generate support documentation from code comments, technical blog posts from design specifications, and API docs from function signatures, you need fewer technical writers and fewer people maintaining these systems. The work becomes more about training the AI systems than doing the work yourself.

Parts of marketing are next. Content creation, email marketing, social media management, competitive intelligence. These aren't the strategic parts of marketing. The AOR (account of record) thinking. The campaign strategy. But the execution layer? The volume production layer? That's where AI gets interesting for marketing.

Engineering support functions are also exposed. QA, testing, some categories of code review, documentation generation. These aren't core engineering work that requires judgment and creativity. These are more systematic tasks where AI can be effective.

What's interesting is the timeline. Support went from 8.3% to 2.88% in 24 months. That's fast. But it's not instantaneous. There's a lag between when AI becomes viable and when companies actually implement it at scale and adjust hiring. There's organizational inertia. There's risk aversion. There's the time it takes to realize the change is real and permanent.

That means sales development hiring probably starts collapsing in the next 12-18 months. Marketing execution hiring maybe 18-24 months. The wave is coming, but it's not all happening at once.

What won't change as dramatically? Functions where the work is fundamentally strategic or creative. Not all engineering, but the core architecture decisions. Not all sales, but the relationship management and large deal strategy. Not all product management. Customer success at the strategic level. Fundraising. Board relations. The work that requires judgment, creativity, and understanding nuance will remain human-driven, at least in the medium term.

The transition is going to create some temporary chaos. Companies will have hiring freezes in areas that are about to be disrupted but haven't fully realized it yet. Other companies will move fast and gain advantage. The talent market will shift. There will be displaced workers who need retraining. This is a significant economic transition.

AI-driven support models show a shift in spending from human salaries to AI systems, with overall costs potentially reducing by 20% (Estimated data).

Why Total Support Spend Might Actually Be Going Up

This seems counterintuitive, but it's important. Companies reducing headcount in support while spending more on support overall. How does that work?

The answer is that "support" spending is moving. It was human salaries before. Now it's software, AI systems, implementation, and training. The budget line changes shape but the total allocation might not shrink.

A company spending

But some companies are actually spending more on support overall because the AI system lets them serve customers better, which drives revenue growth. If a support improvement helps reduce churn by 2%, and you have a

What's definitely happening is that the spending is concentrating. Instead of money spread across 30 support people, it's now concentrated in really good AI systems and a few really good people. The top tier is raising. The middle is being eliminated. This has significant implications for the support industry career path.

Traditionally, support was a career ladder. You started as support rep, you got good at it, you became support team lead, then support manager, then VP of customer success. That path is contracting because the entry-level positions are disappearing. You can't build an organization if there's no entry level.

Some companies are solving this by having people come in through different paths. Technical support becomes a role for junior engineers. Success becomes a role for former sales people. Other companies are offering more training and development for the support roles that remain, paying them more, and making it a premium career path rather than an entry path.

The companies that handle this transition well will be the ones that acknowledge the change and actively manage the career implications. The companies that pretend it's not happening and keep hiring support people the old way will have problems.

The Customer Experience Question

Let's address the elephant in the room. If you're cutting support headcount in half, doesn't customer experience suffer?

The answer is more nuanced than yes or no.

When implemented well, customer experience often improves. Here's why. First, response times get faster. An AI system doesn't have a queue. If you need help at 2 AM on a Saturday, the AI responds instantly. A human support person doesn't work at 2 AM on Saturday. Second, consistency improves. An AI system handles the same types of questions the same way every time. There's no variance based on whether the person is having a good day.

What does suffer is the "someone knows me and my business" feeling. If you've been a customer for three years and have built a relationship with your support person, and then the company cuts back on support staff, you might feel that. The generalized experience is better. The personal experience might be slightly less personal.

The companies that handle this best are explicit about it. They automate the routine stuff so aggressively that when you talk to a human, that human can actually help with complex problems. You get faster routine help and better expert help. The generalist support person is gone, but the specialist is better.

Customer experience is also a function of how well the AI system is trained. A poorly trained AI that gives wrong answers is worse than a slow human. But the leading implementations aren't poorly trained. They're sophisticated systems with ongoing learning and escalation protocols.

The real risk isn't customer experience degradation from the AI itself. The risk is companies going too far, cutting support staff too aggressively, and ending up with neither good AI nor good humans. That's a failure mode you see with any new technology. The sweet spot is usually somewhere in the middle. Enough AI to handle the volume. Enough humans to handle the complexity. Good balance between the two.

What This Means For SaaS Businesses Right Now

If you're running a SaaS company or building one, the implication of the 65% hiring decline is clear: you need to move on support infrastructure now. Not eventually. Now.

First, audit what you're currently spending on support. How many people? What are they doing? How many of those tasks are routine? How many require expertise?

Second, evaluate AI support systems. Not as an experiment. As a real potential replacement for part of your current model. What would it cost to implement? What would the deflection rate likely be? What would happen to your support metrics?

Third, think about restructuring. If you implement AI and it handles 50% of volume, how do you redeploy the support people who are no longer needed for routine work? They could move into success roles. They could move into operations. They could specialize in complex technical support. What makes sense for your business?

Fourth, understand the hiring implications. If you're a fast-growing company that would normally hire lots of support people, you probably don't. You hire fewer support people and you implement good AI instead. If you're a company trying to scale from 50 to 100 people, your support headcount probably grows slower than it would have historically.

Fifth, be honest about the transition. The 65% decline is happening because AI actually works. Pretending it's not real will leave you uncompetitive. Companies that move fast on this get a structural cost advantage. That advantage is real and persistent.

The best companies in your space are probably already thinking about this. If you're not, you're on the wrong side of a significant shift.

AI systems handle approximately 50% of inbound support tickets, effectively reducing human workload and maintaining customer satisfaction. (Estimated data)

The Broader Hiring Trend This Represents

Stepping back from support specifically, the 65% decline is a data point in a much larger story about how AI is reshaping hiring and work.

For the past five years, people have debated whether AI would actually displace workers. The argument usually went something like: "Sure, AI will change jobs, but new jobs will be created, and historically that's how technological transitions work."

That's still probably true at the macro level over 10-20 years. But at the micro level, over the next 2-3 years, there's going to be significant displacement in specific categories. Support is first. Sales development next. Parts of marketing and engineering support after that.

This creates both challenges and opportunities. The challenge is for people whose careers have been built around these functions. The opportunity is for people and companies that can identify where the transition is heading and prepare for it.

Industries that are slow to adopt AI in their hiring will fall behind companies that adopt quickly. This isn't true evenly across all functions, but it's true for routine, high-volume, predictable work. That's a big chunk of what most organizations do.

The companies that figure out how to integrate AI systems into their operations while maintaining good culture, good product, and good customer experiences will thrive. The companies that see AI as a cost-cutting measure and use it that way will probably make short-term gains and long-term mistakes.

Hiring is changing. The 65% decline in support hiring is the proof.

What AI Support Actually Looks Like in Practice

Let's be concrete about what implementing this looks like, because the theory is interesting but the practice is what matters.

Most modern AI support implementations follow a similar pattern. A customer submits a ticket or starts a chat. Instead of being routed to a human support person, the ticket gets to an AI agent first. The AI agent reads the ticket, understands what the customer is asking, and generates a response.

If the AI agent is confident it can resolve the issue, it provides a response. The ticket is closed. The customer is satisfied. Done.

If the AI agent detects that the issue is complex, the ticket gets escalated to a human. The human sees the original ticket plus any analysis the AI did. They have all the context they need to actually help.

What's sophisticated about good implementations is the learning loop. When a human resolves a ticket that was escalated, the system learns from it. The AI system improves. Over time, the deflection rate (percentage handled without human intervention) goes up.

The best systems also have guardrails. They don't try to handle everything. They're conservative about what they claim to have solved. They escalate when they should. They don't make things worse.

Some companies are doing this with off-the-shelf platforms like Intercom's Fin or Zendesk's automation layer. Other companies are building custom systems on top of general AI models. The approach varies, but the pattern is similar.

The deployment process typically takes a few weeks to a few months. You integrate the system. You train it on your knowledge base and past tickets. You test it. You roll it out. You monitor it. You iterate.

The cultural challenge is real though. Support teams that are used to handling tickets directly see an AI system show up and it disrupts their workflow. If the company handles the transition well, the team adapts. If the company is clumsy about it, there's frustration and potential departure.

The best transitions involve support team members in designing the new process. They often become the people who manage the AI system, work on escalations, and improve the training. Instead of disappearing, their role changes. Not everyone accepts that change, but enough do for it to work.

The Economic Pressure Is Relentless

One reason the 65% decline is accelerating (half of it happened in the last three quarters) is that the economic pressure is relentless.

If your competitor implements an AI support system that lets them run support with 50% fewer people, and you don't, you're at a structural cost disadvantage. That's not a hypothetical. That's real. If both of you have

On a typical SaaS operating margin of 20-30%, that $2 million could be the difference between profitable and unprofitable. It could be the difference between spending on product innovation and trying to cut expenses. It compounds.

That's why the hiring collapse is accelerating. Once a few leaders in your space do it, the pressure to follow is enormous. You can't stay competitive without making similar moves.

The pressure is also from investors. VCs are watching this trend. When they see one company reduce support headcount by 50% and maintain or improve customer satisfaction, they're going to ask every other company in their portfolio why they're not doing it. That creates pressure to move faster.

This is how technology disruptions actually work. It's not that everyone adopts at the same pace. It's that adoption accelerates as it becomes clear that the change is real and persistent. Early movers gain advantage. Later movers struggle to catch up. The curve goes from slow to fast to very fast.

We're in the fast phase for support. You're going to see the decline steepen further.

In scaling companies, the number of generalist support roles is decreasing, while specialist roles and AI automation are increasing. Estimated data based on industry trends.

Preparing Your Organization For What's Coming

If you're in a function that might be disrupted over the next 2-3 years (sales development, support, parts of marketing or engineering), what do you do?

First, don't panic. The disruption is real, but it's not overnight. You have time to prepare. Sales development hiring is probably going to start declining noticeably in 12-18 months. That gives you time to upskill, pivot, or make strategic moves.

Second, understand which parts of your work are vulnerable. The routine, high-volume, predictable parts are most at risk. The strategic, creative, complex parts are less at risk. If you're spending 80% of your time on routine work and 20% on strategic work, that ratio is probably going to flip. The question is what you do about it.

Third, develop skills that complement AI rather than compete with it. In support, that means developing deeper technical skills, customer relationship skills, or domain expertise. In sales development, that means developing research skills, strategic thinking, or relationship building. The work that remains will reward people who can do things AI can't do yet.

Fourth, pay attention to which companies are handling the transition well and which are mismanaging it. The companies that are thoughtful about integrating AI while maintaining culture will be good places to work. The companies that treat it as a pure cost-cutting measure often create chaos and dysfunction. If you're looking for work, work for the thoughtful ones.

Fifth, think about your career architecture. If the entry-level role in your function is disappearing, what's your path forward? Some people will upskill into the specialist roles. Some will move into adjacent areas. Some will move into roles managing AI systems. The paths are there, but they're different than they would have been two years ago.

The Work That Remains Will Be Better

Here's something that often gets lost in discussions of AI disruption: the work that remains for humans is often better.

Generalist support reps do a lot of boring, repetitive work. Password resets. Billing questions. "How do I do X" questions. It's necessary work, but it's not particularly fulfilling.

When AI handles that work, the remaining support work becomes solving interesting problems. Understanding why a customer is stuck and helping them get unstuck. Building relationships with important customers. Understanding deeper patterns in how customers use your product. That's more interesting work.

The same dynamic plays out in other functions. Sales development reps do a lot of research and outreach. When AI handles research and initial outreach, the remaining work is building relationships with prospects who are actually interested. That's more interesting.

This doesn't mean the transition is painless. It means that on the other side of the transition, the work is often better. The people who can make the transition will find more fulfilling careers ahead of them.

The risk is that organizations don't manage the transition thoughtfully. They cut headcount too aggressively. The remaining people are overwhelmed. The work that's supposed to be better becomes worse. That's a failure mode that plenty of organizations will hit.

The organizations that do it well—that actually restructure roles to be better, that invest in training, that pay well for the work that remains—will have higher engagement, lower turnover, and better outcomes.

The Competitive Moat This Creates

Here's something interesting from a strategy perspective. Companies that move fast on AI support infrastructure and do it well create a competitive moat.

It's not just that they have lower costs. It's that they have a better, faster customer support experience. They have happier customers. They have happier employees because the work is less drudgery. They have better data about how customers use the product because the AI system creates insights humans would miss.

Over time, those advantages compound. Lower churn. Higher NPS. Better product decisions informed by support data. Lower operational cost. That's a powerful combination.

Companies that lag on this will find it harder and harder to catch up. The moat gets wider. The companies ahead keep getting further ahead.

This is why the 65% hiring decline is accelerating and will continue to accelerate. It's not just about cost. It's about building structural advantages that persist.

If you're a founder or executive, this should be in your strategic thinking right now. Not as something that might happen someday. As something that's happening to your competitors right now.

AI systems handle approximately 60% of support tickets, with humans managing 20% and another 20% being complex cases that require deeper human involvement. (Estimated data)

The Counterargument (And Why It's Mostly Wrong)

There's a reasonable counterargument that goes something like: "AI in support isn't really that good. Companies are cutting support staff for the wrong reasons. Customer experience is going to suffer, and we'll see the consequences in churn and negative reviews."

There's a grain of truth in this. Some companies will implement AI poorly. Some customers will have bad experiences. There will be edge cases where the AI creates problems.

But the broad argument doesn't hold up against the evidence. Companies are implementing AI support at massive scale. They're not doing it as experiments. They're doing it in production, at real companies, with real customers. The deflection rates are good. Customer satisfaction metrics are holding up or improving. The data just doesn't support the "AI is bad at support" narrative.

What's probably true is that AI is bad at certain types of support. Highly nuanced emotional situations. Unique edge cases. Technical issues that require deep debugging. But those aren't 60% of support volume. They're maybe 20%. AI doesn't need to be perfect. It just needs to be good enough at the common cases and willing to escalate on the hard cases.

The economic pressure is also irresistible. Even if AI support were mediocre (which it mostly isn't), companies would still adopt it because the cost savings are so large. The fact that it's actually good at the work just accelerates adoption.

So the counterargument exists, but it's not particularly compelling in the face of the evidence. The AI works. The hiring decline is real. This is happening.

What The Next 18 Months Look Like

If you want a practical timeline, here's what's likely to happen in the next 18 months based on the trajectory we're seeing.

First 6 months: More companies implement AI support systems, especially companies scaling from

Months 6-12: Sales development hiring starts to noticeably decline as companies realize the same logic applies. Companies start implementing AI for sales research and initial outreach. The hiring curves for sales development and marketing execution start flattening. More companies merge support and success functions as the structural lines blur.

Months 12-18: The shift becomes undeniable. Every earnings call has a question about how AI is affecting headcount plans. Every SaaS company has taken a position on support AI and made a decision about their approach. Some companies have moved fast and have structural advantages. Others are catching up. The labor market for support and sales development reps is noticeably tighter because demand dropped faster than supply adjusted.

By month 18, this becomes the normal way companies think about support. The conversation shifts from "should we implement AI support" to "how do we optimize our AI support implementation." That's when you know the transition is real.

The Talent Challenge This Creates

One of the less discussed implications of the 65% hiring decline is the talent problem it creates.

FOR the next few years, if you're a company that wants to hire good support or sales development people, you'll be able to hire them at reasonable costs because there's suddenly a lot of supply (people whose companies cut support headcount and they need jobs). But after a few years, the supply dries up because there's no more entry-level hiring happening.

That creates a labor market challenge. You have a glut of experienced support people who need jobs. You have a shortage of junior support people coming through because the entry-level jobs are gone. You have an aging population of support professionals with nowhere to grow.

Some companies will solve this by training people from other backgrounds. You hire ambitious junior engineers and train them into support roles. You hire smart generalists and develop them into support specialists. But that takes intentional effort.

The companies that don't handle this well will find themselves short on talent in a few years. The ones that do will have advantages.

From a broader societal perspective, this is why education and training in technology skills matters. The support jobs are going away, but the need for people who understand customer problems and can bridge customer needs with technical solutions isn't going away. It's just the path to that career is different.

Building Your Support Strategy In 2025

If you're building a support organization or restructuring one right now, here are the key decisions to make.

First, how much volume are you expecting? Based on that, what's the right balance between AI and human resources? The math is different for different sized companies. A company doing

Second, what types of support do you need? If your product is technically simple, AI can handle more. If it's technically complex, you need more humans. If you have a specialized customer base with unique needs, you need different skills.

Third, what's your customer satisfaction baseline? What do your customers expect? Support requirements vary enormously by industry, customer type, and price point. Enterprise customers expect higher touch. SMB customers often prefer self-serve AI. Match your approach to expectations.

Fourth, how will you train your AI system? This is underestimated by a lot of companies. The AI needs good data to work with. You need a knowledge base. You need documentation. You need feedback loops. Companies that invest in this properly get much better results than companies that deploy hastily.

Fifth, how will you handle the human side? Are you restructuring existing support teams into new roles? Are you hiring new people to manage the AI system? Are you combining support and success? Be clear about this and communicate it.

Sixth, what metrics matter to you? Deflection rate? Response time? Customer satisfaction? Cost per ticket? Different companies optimize for different things. Be clear about what success looks like.

Getting these decisions right is the difference between a transformation that goes well and one that's chaotic.

FAQ

What exactly is the 65% hiring decline in customer support?

The 65% decline refers to customer support roles dropping from 8.3% of all new hires in Q4 2023 to just 2.88% in Q3 2025. This data comes from Pave's analysis of 386,500 new hires across companies. It means that when companies hire new people, a much smaller percentage of those hires are going into customer support roles compared to two years ago. The decline is accelerating, with almost half the total drop happening in just the last three quarters.

Why is customer support hiring declining if companies still need customer support?

Companies haven't stopped providing customer support, they've restructured how it works. AI systems now handle 40-60% of support volume that humans used to handle. Instead of hiring 10 new support generalists to handle growth, companies implement an AI system and hire 2 highly specialized support specialists. The total headcount goes down, but the support capability often stays the same or improves. Revenue and company growth have continued even as support headcount has shrunk dramatically.

Is customer experience actually suffering as a result of this?

At well-implemented deployments, customer experience typically improves because AI handles routine issues instantly while humans focus on complex problems. Customers get faster responses to standard questions (password resets, basic troubleshooting) and better expertise when they need escalated help. However, the experience can suffer if companies cut support staff too aggressively without adequate AI implementation. The sweet spot is usually having good AI deflecting routine work while maintaining enough experienced humans for complex cases.

Which other job categories will experience similar hiring declines to customer support?

Sales development is next, likely declining noticeably in 12-18 months as AI takes over research and initial outreach. Parts of marketing (content creation, email marketing) are vulnerable because execution-layer work is predictable. Technical support roles, testing, and code review functions are also exposed. Functions that are strategic, creative, or require deep judgment (core engineering, relationship sales, product strategy) are more protected. The pattern is that high-volume, predictable, routine work faces disruption first.

What does restructured support look like at companies doing this well?

Well-executed restructuring typically involves AI handling tier-one support (password resets, basic questions, standard troubleshooting), with humans handling tier-two escalations and complex issues. Remaining support roles become more specialized, better-paid (often

Is this a temporary efficiency trend or a permanent structural change?

This is a permanent structural change. Companies aren't trying AI support for three months and reverting to humans. They're fundamentally redesigning how support works based on what AI can do. This shows up in hiring patterns (65% decline) and in how vendors like Zendesk and Intercom are repositioning their entire products around AI. The change is real, persistent, and accelerating. Once a company restructures support around AI and sees the results (lower costs, maintained or improved satisfaction), they don't go back.

How should people in support or sales development roles prepare for this disruption?

Focus on developing skills that complement AI rather than compete with it. In support, that means technical depth, customer relationship management, and domain expertise. In sales development, it means research skills, strategic thinking, and relationship building. Understand which parts of your current work are routine (vulnerable to AI) and which are strategic (safer). Look for organizations handling the transition thoughtfully rather than treating it purely as cost-cutting. Consider upskilling into adjacent areas that might be more stable. The work that remains will be more interesting and better-paid, but the path there requires intentional preparation.

What happens to customers of support software like Zendesk and Intercom if so much support work goes away?

These companies are thriving despite support hiring declining because they've shifted to AI-powered platforms. Zendesk's traditional helpdesk business is flat, but their AI automation business is growing rapidly. Intercom went all-in on Fin (their AI support agent) and that's becoming their primary product. Gorgias' AI agent business is growing at 100% annually near $100M ARR. The platform vendors are winning because companies still need to manage support, they just need AI-powered systems instead of hiring more humans.

Will support become a dead-end career path or will new opportunities emerge?

Entry-level support roles are declining, which creates a real challenge for career paths. However, specialist support roles that remain are better-paid, more interesting, and increasingly connected to success and product strategy. New opportunities are emerging in managing AI systems, analyzing support data for product insights, and bridging customer needs with technical solutions. The path to a support career is changing from "entry-level rep to manager" to "specialized expert," but meaningful careers will persist for people who upskill and adapt.

The Real Opportunity Here

The biggest opportunity in this shift isn't for companies cutting support costs. It's for companies that see this transition and use it to build better customer relationships.

When you implement AI support well, you free up humans to do the relationship work that actually drives retention and expansion. You get better data about how customers use your product. You have happier support people doing more interesting work. You have faster response times for routine issues and better expertise for complex ones.

The companies that gain the most from this transition are the ones that treat it as a chance to improve customer experience, not just cut costs. The companies that see it as pure cost-cutting often end up with worse outcomes.

The data is clear. The 65% decline in hiring is real. It's accelerating. It's not temporary. But how you respond to it determines whether it's a strategic advantage or a competitive liability.

Move thoughtfully. Communicate clearly. Invest in good systems. Treat the people well. Do that, and you come out of this transition stronger. Rush it, and you'll regret it.

Key Takeaways

• Customer support hiring dropped 65% in 24 months (8.3% to 2.88% of new hires) because AI now handles 40-60% of ticket volume
• Companies aren't eliminating support quality—they're restructuring it with AI handling tier-one routine issues while specialists handle complex escalations
• Support roles are shifting from
  $50-65K generalists to$
  90-120K specialists focused on customer success and strategic relationships
• Sales development and marketing execution roles will experience similar hiring declines in the next 12-24 months following the same AI disruption pattern
• This is a permanent structural change, not a temporary efficiency trend; companies that move fast gain compounding competitive advantages

Related Articles

• Chrome's Auto Browse Agent Tested: Real Results [2025]
• Who Owns Your Company's AI Layer? Enterprise Architecture Strategy [2025]
• How AI Transforms Startup Economics: Enterprise Agents & Cost Reduction [2025]
• Meridian AI's $17M Raise: Redefining Agentic Financial Modeling [2025]
• OpenAI's Responses API: Agent Skills and Terminal Shell [2025]
• Observational Memory: How AI Agents Cut Costs 10x vs RAG [2025]