Kief Studio

Kaseya Promised Managed Service Providers 5x Capacity Without Hiring. Here Is What That Exposes.

Build

  • Published
  • Updated
  • 10 min read
Violet-magenta and soft gold light trails sweep in from the edges and meet at a bright point in the centre of a black field.

TL;DR: In April 2026 Kaseya launched an agentic IT management platform and pitched it to managed service providers as more capacity without matching hires. Automating routine tickets is useful. But once the commodity work runs itself, clients see what is left. Point the freed hours at integration, governance and builds that clients cannot buy off a shelf.

Update, October 2026: an earlier version of this note leaned on forum sentiment scores, layoff and debt figures, industry deal counts and margin numbers we could not trace to a primary source. We rebuilt it around Kaseya's own announcement, government data and guidance, and analyst statements we could check. The angle is the same.

If you run one of the many managed service providers that live on help desk tickets, you probably heard the pitch at the end of April. Kaseya, one of the larger software vendors to managed service providers, announced what it calls the first agentic IT management platform. Trade coverage summed up the promise as up to 5x more capacity without a matching increase in hiring.

If your labor costs keep climbing and good technicians are hard to find, that sounds like relief. I think it is something else first: an exposure event. When the routine work gets cheap for everyone, your clients start asking what they are paying you for. Here is what was actually announced, what the public data says about where support work is heading, and where I would point the capacity.

What did Kaseya actually announce for managed service providers?

Kaseya's press release, dated April 28, 2026, says the platform is "built on more than 1 billion help desk tickets, 3 exabytes of backup data, and 17 million managed endpoints." The first agent, which Kaseya calls a Digital Specialist, handles ticket triage, which "automatically categorizes and routes tickets with high accuracy." Kaseya says the "Ticket Triage Digital Specialist is generally available today for Autotask Ultimate customers," with more specialists to follow. (Kaseya via PR Newswire (opens in a new tab))

The release also says the platform is different from tools "that surface insights for humans to act on," because it acts "across systems instead of bouncing alerts back to a human." (Kaseya via PR Newswire (opens in a new tab)) That is the real shift. Earlier AI features suggested. This one is meant to do.

What is verified, and what is a vendor claim?

Read the numbers carefully. The dataset size is Kaseya describing its own platform. The accuracy figure comes from an early customer, who said "20 to 30% of our tickets aren't categorized correctly today" and that "we believe the Ticket Triage Digital Specialist will eliminate up to 80% of those errors." (Help Net Security (opens in a new tab)) That is a belief from a test, not a measured result.

The capacity headline is the same. The press release itself does not use the 5x figure. Trade coverage reported that "Kaseya claims IT teams can achieve up to 5x more capacity" (TechNewsHub (opens in a new tab)), and another outlet described the aim as "allowing partners to scale their operations without increasing headcount." (MSP Success (opens in a new tab)) Nobody has published an independent test of it yet. Treat it as a sales target, not a benchmark.

None of that makes the product bad for managed service providers. Ticket triage is tedious, error-prone work, and a tool that routes tickets correctly saves real time. My point is about what happens next.

5x capacity to do what?

That is the question I did not hear anyone ask on stage.

If the extra capacity goes into more password resets, more printer tickets and more "have you tried restarting it," you are not five times more valuable. You are doing the same work faster than before, on the same platform your competitors just bought. Speed on commodity work is not a moat when the vendor sells the same speed to all managed service providers at once.

The other risk is quieter. Clients pay managed service providers partly for the feeling that someone is handling it. When a client sees their tickets close in minutes with no person involved, some will ask a fair question at renewal: what is the rest of the invoice for? If the honest answer is "mostly the same tickets," that conversation goes badly.

Where is help desk work heading, according to the government data?

The Bureau of Labor Statistics already expects automation to shrink routine support work. Computer support specialists held 903,100 jobs, with median pay of $62,890 a year in May 2025. Employment "is projected to decline 3 percent from 2025 to 2035." (U.S. Bureau of Labor Statistics (opens in a new tab))

The reasons BLS gives line up with this announcement almost word for word. User support jobs decline "as organizations continue to implement automated tools, such as chatbots, for troubleshooting." And on the hopeful side: "This use of automation may free up some computer user support specialists to handle more complex cases." BLS also expects smaller organizations to keep choosing "to contract with outside firms for these services rather than to employ computer network support specialists." (U.S. Bureau of Labor Statistics (opens in a new tab))

Put those two facts side by side. Small businesses will keep buying support from outside firms. And the simple part of that support is the part being automated. So the managed service providers that keep their clients will be the ones selling the complex cases.

What is left for managed service providers when tickets triage themselves?

A hand lifting a desk phone handset in a dim office, the moment an automated help desk hands a call to a technician
When triage runs itself, the calls that reach a technician are the hard ones. That is where managed service providers earn their fee.

The work that does not automate well is the work that depends on knowing one client's business. Why this accounting firm's document system breaks every tax season. Which manufacturer's ERP export has to land in the shipping tool by 6am. What a dental office is allowed to store where. That is judgment, history and integration, and it is the part clients will still pay for when the routine tickets cost close to nothing.

Why does agent trust matter more than agent speed?

An agent that acts in a client environment is a new kind of privileged account. The federal guidance for managed service providers already treats your access as the risk to manage, and an agent inherits that access.

A 2022 joint advisory from CISA, the NSA, the FBI and international partners notes that managed service providers "provide services that usually require both trusted network connectivity and privileged access" to customer systems. It tells providers and customers to "Ensure MSP-customer contracts transparently identify ownership of ICT security roles and responsibilities," to "Enforce MFA on MSP accounts that access the customer environment," and says customers should ensure that their MSP applies least privilege "to both provider and customer network environments." (CISA joint advisory AA22-131A (opens in a new tab))

Now picture an agent doing remediation across fifty client tenants. Every one of those principles still applies to managed service providers running agents, and a good client will ask how. We cover the security side in our note on agentic security.

Questions to answer before an agent touches a client environment

  • What can it change? Read access, ticket updates, or real actions like resetting accounts and pushing scripts. Write it down per client.
  • Whose identity does it act as? A named service account with the least privilege the task needs, not a shared admin login.
  • Where is the log? Every action the agent takes should be visible to you and, on request, to the client.
  • Who approves the risky steps? Decide which actions need a person to say yes first.
  • What does the contract say? If an agent makes a bad change at 2am, the client agreement should already say who owns that.

The NIST AI Risk Management Framework (opens in a new tab) is a good free starting point here. NIST says it "is intended for voluntary use and to improve the ability to incorporate trustworthiness considerations" into how AI is designed and used. It gives you a vocabulary clients and insurers already recognize.

Why do so many agentic AI projects fail?

Gartner predicted in June 2025 that "over 40% of agentic AI projects will be canceled by the end of 2027, due to escalating costs, unclear business value or inadequate risk controls." (Gartner (opens in a new tab)) Gartner also named a trend it calls "agent washing," older chatbots and automation relabeled as agents, and estimated that of thousands of vendors claiming agentic solutions, "only around 130 offer real agentic features." (MarTech (opens in a new tab))

Look at the three reasons Gartner gives. Cost, unclear value and weak risk controls are not model problems. They are scoping, integration and governance problems. A platform vendor can ship the agent to thousands of managed service providers. It cannot know which of your client's workflows should never be automated, or how that client's data actually flows.

Rushing agents into client work

Most failed rollouts I see skip the boring part. Nobody wrote down the workflow, the exceptions, or who checks the output. The agent works in the demo and then meets the client's real systems: the custom field nobody documented, the vendor portal with no API, the approval that lives in someone's inbox. That is where projects stall, and where the cost overruns come from.

Why is integration the work that holds its value?

Agent features will spread across every platform vendor that sells to managed service providers, and they will get cheaper. Integration does not ride that curve down, because it is specific. Wiring automation into one company's processes, exceptions and history takes someone who understands both the technology and that business.

What integration work actually involves

  1. Mapping the workflow. Every step, every system, every hand-off, including the manual ones nobody admits to.
  2. Agreeing access. Which systems may connect, and what data each one is allowed to see.
  3. Building on authorized interfaces. Real API integration, with retries and clear error handling, not screen scraping.
  4. Checking outputs. Quality gates that catch a wrong answer before it reaches a customer, with a person reviewing what fails.
  5. Watching it run. Monitoring so a broken sync raises a flag instead of losing data quietly.

None of that comes in the box. It is project work with real engineering, and clients pay for it because it is theirs.

What should managed service providers do with the freed capacity?

Here is how I would plan the first 90 days if I ran one of these managed service providers and had just switched on ticket automation.

A 90-day plan

  1. Measure before and after. Pull your ticket counts by category and your average handling time now. If the vendor's claim holds for you, you will see it. If not, you will know that too.
  2. Write an agent policy per client. Use the five questions above. Put the answers in the client file and the contract.
  3. Pick three clients with a painful workflow. The double data entry, the report someone builds by hand every Monday. Offer to fix one as a project.
  4. Decide build or partner. If you do not have engineers who build integrations, partner for it and keep the client relationship. Many managed service providers already handle custom development or penetration testing this way.
  5. Change the renewal conversation. Report on what you built and what you prevented, not only on tickets closed.

The managed service providers that come out of this stronger will be the ones who took the hours agents freed up and pointed them at work that requires engineering, not only platform administration. The ones that do not will find that more capacity for commodity work just makes them easier to replace when the next vendor offers the same thing for less.

Frequently asked questions

What did Kaseya launch in April 2026?

An agentic IT management platform powered by what it calls Kaseya Intelligence. The first agent, Ticket Triage, categorizes and routes tickets and became generally available for Autotask Ultimate customers.

Is the 5x capacity figure proven?

No. It is Kaseya's claim as reported in trade coverage, and the press release does not include it. No independent test has been published. Measure your own ticket data before and after.

Will AI agents replace help desk technicians?

BLS projects computer support specialist employment to decline 3 percent from 2025 to 2035, partly because of automated tools like chatbots. It also expects automation to free some specialists for more complex cases.

What should managed service providers do with the time agents save?

Point it at work clients cannot buy off a shelf: integrations, governance, custom builds and projects that fix painful workflows. That work depends on knowing the client.

How should an MSP govern an AI agent in client environments?

Treat it as a privileged account. Least privilege, a named identity, full logging, approval for risky steps and clear contract terms, following CISA's MSP guidance and the NIST AI Risk Management Framework.

Can an MSP offer integration work without hiring engineers?

Yes. A build partner can do the engineering under your brand and contracts while you keep the client relationship, the way many providers already handle custom development.

Where we fit

Kief Studio does white-label development for agencies and studios, and for partners who need dependable engineering behind their brand. We build it, your name stays on it, and we stay out of the client relationship unless you bring us in. The fit is custom web applications, integrations and the infrastructure behind them: the work agentic platforms create demand for but do not deliver.

On the integration side, our API integration work connects the tools your clients already pay for, built on authorized APIs only, hosted and watched so a failed sync gets noticed. When an AI workflow is part of the job, our AI implementation work adds error handling, quality gates and a human path for anything the model is unsure about. If agents just freed up your bench and you want to sell more than faster tickets, that is the conversation to have.

First conversation is free. No commitment.

Tell us what your clients are asking for

Related articles

  1. Support

    Your Clients Want AI Services. Most MSPs Can't Deliver Them Yet.

    Clients want AI services from their IT provider. What Census, GTIA, MIT and Gartner data show about the gap, and how an MSP can deliver AI services safely.

    Brian Gagne

  2. Support

    Your Clients Want AI Services. AI Service Delivery Is the Support Ticket After the Demo.

    AI service delivery fails after the demo, not during it. What Klarna, Air Canada and research show, and support terms to write before you sell an AI agent.

    Amelia Gagne

  3. Build

    Should Your Agency White-Label a Core Service This Year? Run the Numbers First

    Should your agency white label a core service? What changed in 2026, where white label fails, and a one-sheet test to decide between a partner and a hire.

    Brian Gagne