AI Is Reshaping the MSP Staffing Model, Not Shrinking It
Randy Hall, CEO
AI is not cutting headcount inside MSPs, it is rearranging who gets hired and what they do all day. Kaseya's 2026 State of the MSP Report found that MSPs citing difficulty finding and hiring skilled technicians jumped from 9% to 16% year over year. Automation, meanwhile, is absorbing routine tickets. The business model shift is not fewer people, it is a different mix of people and a different ratio of clients each one can carry.
Is AI actually replacing MSP technicians?
Not according to the people running MSPs. In Kaseya's 2026 State of the MSP Report, 21% of MSPs expect AI to reduce headcount, while 59% expect it to eliminate tedious tasks and 44% expect it to provide more time to strategize and grow the business. That is a workforce being redeployed, not eliminated. The report's own summary is that AI is seen as supporting technicians, not displacing them.
That distinction matters for how you plan the next four quarters. A headcount freeze built around "AI will replace people" is planning for an outcome most owners in the same survey do not expect. A headcount plan built around "AI will change what people do" is planning for the outcome the data actually points to, and it changes which roles you post, not how many people you keep on payroll.
Why is hiring skilled technicians getting harder while AI absorbs routine work?
Because the routine work was never the hard part to staff for. Kaseya's report found the share of MSPs citing difficulty finding and hiring skilled technicians jumped from 9% to 16% in a single year. AI is closing the gap on ticket volume, but it does not close the gap on judgment, on the engineer who reads a strange log line and knows it is a precursor to a ransomware event. Automation removed the easier tickets that used to buy a junior technician time to grow into that judgment on the job.
The ratio that actually changes: clients per senior engineer
The staffing model MSPs built for the last two decades assumed a pyramid: a wide base of tier 1 technicians answering tickets, a smaller group of tier 2 engineers escalating the hard cases, and a thin layer of senior staff who rarely touched a ticket queue at all. AI collapses the base of that pyramid first, because triage, password resets, and routine patch confirmation are exactly the tasks large language models and scripted remediation handle well. What does not collapse is the senior layer. A smaller number of senior engineers, now supervising automated remediation instead of doing it by hand, can reasonably cover more accounts than before, which changes the ratio your pricing and staffing plans were built around.
That ratio, not total headcount, is the number worth tracking every quarter. An MSP that keeps hiring tier 1 seats to match client growth is paying salary for work AI already does for free. An MSP that holds tier 1 flat and grows senior capacity per account is the one whose margin improves as it scales.
What top performing MSPs are hiring for right now
ScalePad's 2026 MSP Trends Report found that top performers are increasing recruitment and hiring, with a focus on sales, security specialists, and general technicians. Demand for AI skills points the same way. CompTIA's State of the Tech Workforce 2026 report found more than 275,000 active job postings in January 2026 that referenced a need for some level of AI skills.
Those postings include both dedicated AI positions and jobs that require AI fluency on top of a traditional role. If your recruiting pipeline still looks like 2019, it is worth checking whether it matches the roles above.
The tier system built for humans does not fit an AI-augmented shop
Most MSPs still run an org chart shaped for a world where every ticket needed a human hand at each tier. That structure gets expensive fast once AI is doing tier 1 work whether you plan for it or not, because you end up paying junior salaries for judgment work that has actually gotten scarce, not for ticket volume. A more honest staffing model looks like this.
| Role in the old tier model | What changes under AI |
|---|---|
| Tier 1: ticket intake and triage | Largely automated, headcount need drops fastest |
| Tier 2: escalation and troubleshooting | Shrinks in size, shifts toward supervising automated remediation |
| Tier 3: senior engineering and architecture | Grows in relative importance, each engineer covers more accounts |
| Security and compliance specialist | New hiring priority, driven directly by client demand |
None of this means you fire your way to a leaner org chart overnight. It means the next hire should close a judgment gap or a skill gap you can name, not fill a seat shaped like last year's ticket queue. Training existing staff into those judgment and security roles is often faster and cheaper than hiring cold, which is the case for structured technician certification and staffing pathways built for this shift rather than waiting on a tight labor market.
What this does to your margin and your renewal conversations
The staffing shift only matters if it shows up in numbers your clients feel. Kaseya's report found 48% of MSPs say AI and automation will be the top client need, ahead of security at 42% and backup at 36%, yet only 13% of MSPs have turned automation and AI into a meaningful revenue stream. That gap is a staffing problem as much as a sales problem. Without the security specialist and senior engineering capacity clients are asking for, you cannot sell what they already want, and a client who cannot get AI-driven service from you will find someone who staffed for it first.
Retention follows the same logic. A client who hears "we're working on our AI roadmap" at renewal time is hearing that you have not yet made the staffing change this data says other MSPs already started. I expect the compliance reviews and insurance renewals your clients sit through to start asking not just whether an AI tool flagged an anomaly, but who on your team reviewed it and signed off. That is a staffing question worth having an answer to before a client's auditor asks it, not after.
There is a talent cost to waiting, too. Hiring difficulty is climbing, and the clock on sourcing a skilled technician does not start until you have already decided what role you are hiring for. Owners who wait for a client escalation to force the decision are starting that clock late, on a role they had to define under pressure instead of on their own timeline.
Where to start if your org chart has not moved yet
Figuring out which roles to shift first does not require guesswork. Running your current stack and client mix through a tool built to show where automation already covers work your org chart still staffs for takes an afternoon, not a quarter, and it tells you which tier to shrink before you post a job listing for it. The platforms that let a smaller senior team cover more accounts without dropping response times already exist, and the full catalog is worth reviewing before you assume headcount is the only lever you have.
The staffing model that wins the next two years is smaller at the base, deeper at the top, and built around tools that make judgment scale instead of headcount. See the full stack and decide which tier to invest in next.