I do a prediction series to close out the end of the year. It involves not only a series of predictions for the coming year, but also a look back at my predictions for this year and an accountability of what I predicted. So now, let’s take a look at how my 2025 predictions actually played out. This is the Business of Tech.
Let’s start with this one.
Prediction: Generative AI will solve the reporting problem and become the primary way businesses extract value from their data.
Outcome? I’m calling this one partially accurate. There’s no question that generative AI has changed how people interact with data. Natural language querying, conversational analytics, and AI-generated summaries are everywhere in business intelligence tools now. But did it solve the reporting problem? Not entirely. Data quality, governance, and trust in AI-generated outputs are still real issues. Generative AI improved access to insights, but it didn’t magically fix broken data pipelines.
Prediction: Autonomous IT systems will reach maturity and see widespread SMB adoption.
Outcome? This is another partial hit. We’re absolutely seeing more automation, self-healing systems, and predictive maintenance in MSP tools. AI-driven remediation is real. But “widespread SMB adoption” might be a stretch. Most SMBs still run on enhanced RMM and automation—not fully autonomous systems. The tech is maturing, but broad market maturity? Not quite there yet.
Prediction: Legal AI services will become a valuable offering for small businesses, packaged by providers.
Outcome? Mixed result. AI-powered legal tools absolutely took off for contract review, document drafting, and compliance checking. Small businesses are using them. But MSPs didn’t broadly become legal service providers. What they did embrace instead was AI compliance, data governance, and regulatory advisory. So the legal AI part hit—but the MSP packaging angle only partially landed.
Prediction: Decentralized MSP models leveraging blockchain and peer-to-peer systems will enter the mainstream.
Outcome? Missed this one. While decentralized infrastructure and blockchain-based storage continue to evolve in labs and niche platforms, they did not enter the mainstream MSP model in 2025. Traditional centralized platforms still dominate backup, authentication, and disaster recovery in the MSP market. This stayed conceptual.
Prediction: AI co-governance frameworks will shape MSP services.
Outcome? Accurate. Governments and global bodies absolutely moved forward with structured AI governance—risk-based AI regulation, ethical frameworks, and compliance models are now real forces. And MSPs are responding with AI governance offerings, policy development, audits, and compliance services. This one landed squarely in the real world.
Prediction: Enhanced AI adoption services—AI readiness assessments and adoption strategy—will take off.
Outcome? This one hit. In 2025, many providers now openly market AI readiness assessments, pilot programs, and adoption consulting. SMBs are no longer “should we look at AI?”—they’re asking “how do we start responsibly?” And MSPs are clearly stepping into that advisor role.
Prediction: MSPs will begin evolving toward strategic outcome providers.
Outcome? Accurate. The shift is very real. Providers are increasingly measured not just on uptime and patching, but on business outcomes—productivity, resilience, security posture, and operational efficiency. AI and automation are handling more backend work, while human expertise is moving up the value stack into consulting and strategic guidance.
Prediction: Outcome-based pricing models will emerge, potentially using smart contracts.
Outcome? Partially accurate. Outcome-based pricing definitely gained traction as a concept—especially in consulting-heavy engagements. But blockchain-enabled smart contracts? That part mostly stayed theoretical. Most MSPs experimenting with outcome pricing are still using traditional contracts and performance metrics, not Web3 infrastructure.
Prediction: Fraud prevention will become a standard MSP offering.
Outcome? Mixed result. Fraud risk exploded in 2025—account takeovers, payment fraud, and impersonation attacks surged across industries. MSPs absolutely expanded phishing protection, identity security, and transaction monitoring in some verticals. But calling fraud prevention a universal, standard MSP service across all SMBs? That’s not quite supported yet. It’s growing, but unevenly.
So what’s the scorecard?
A few clean hits:
- AI governance shaping services
- AI readiness and adoption consulting
- The shift toward outcome-driven MSP models
A few partial wins:
- Generative AI in reporting
- Autonomous IT systems
- Legal AI services
- Outcome-based pricing
- Fraud prevention
And one clear miss:
- Decentralized, blockchain-driven MSP models entering the mainstream
Now, why do we care?
I make these predictions less to be right—although I do enjoy the accountability—and more to stimulate your thinking as you build your strategy. The biggest takeaway from 2025 isn’t any single technology. It’s that AI governance, business outcomes, and advisory services are now central to what it means to be a modern provider.
And in reviewing these, a pattern stands out. The predictions that landed were all about governance, outcomes, or advisory work — the places where providers actually influence business value. The partials tended to be technologies that matured unevenly. And the miss, the decentralized MSP model, is a good reminder that adoption curves don’t accelerate just because the tech is interesting. Real change happens when operations, economics, and risk all line up.
Because the big story going into 2026 isn’t “more AI.” It’s that automation is becoming the production system of IT services, identity is becoming the control layer for machines, and vendors are going to be judged as much on automation safety as on features. And as that happens, the economics start to split more clearly between labor-scaled and automation-scaled MSPs. That’s the backdrop for the predictions I’m about to make.
So these are my 2026 predictions, and a few of these are intentionally uncomfortable.
First prediction: If your MSP revenue still scales with human labor in 2026, your model is already under pressure.
I don’t think “agents” are the story. I think automation is the story. This is the steady extension of RPA, orchestration, remediation, and workflow automation—with AI inside it.
That shift creates pressure to rethink how a services business actually scales. If your model still relies on people babysitting automations or handling work that should already be systemized, you’re going to feel it. The providers who get ahead will be the ones increasing their Managed Endpoints Per Technician by doing three things: identifying where humans are still in the loop unnecessarily, defining outcomes in concrete terms rather than generic uptime metrics, and putting real governance around how automation runs in production. That’s how you move from a labor-scaled business to an automation-scaled one, and it’s the divide that starts to widen in 2026.
Second prediction: Someone is going to be personally named in an automation-related lawsuit.
Not “the system caused harm.”
Not “the vendor failed.”
A human.
Courts are not going to accept “the machine did it.” If an automated process causes damage—financial, regulatory, operational—someone will be accountable. A big part of that accountability is identity. Every script, every agent, every automation process now needs its own Non-Human Identity. The risk in 2026 isn’t just compromised users — it’s over-permissioned automations being hijacked. The same least-privilege principles you apply to people now have to apply to machines.
Third prediction: Infrastructure management is no longer where the premium lives.
Patching, backups, remediation, scaling—more and more of that becomes expected, automated, and assumed. That doesn’t mean it goes away. It just means customers stop thinking of it as differentiated.
If “we manage your systems” is still the center of your value proposition in 2026, I think you’ll feel increasing pressure to explain why that’s worth more than table stakes. Think of it this way: your clients expect the lights to turn on. They don’t pay extra for it. In 2026, automated patching and self-healing systems are the lights. The premium is found in the Proactive Advisory—telling them why they need a new cloud architecture, or how their AI strategy impacts their regulatory footprint. Stop focusing your sales conversation on the things that are now being handled by automated systems.
Fourth prediction: Automation governance becomes more important than the tools themselves.
Most providers will have access to the same categories of platforms. The differentiator won’t be which RMM or which security stack you picked.
It will be things like:
- How safely you automate
- How you handle exceptions
- How tightly you control change
- How well you can audit what actually ran
And the vendors who don’t provide real guardrails — auditability, identity chaining, rollback — won’t just be inconvenient. They’ll be liabilities. In 2026, MSPs will choose tools based on the safety of automation, not just the speed of it.
The providers who get really good at governing automation, not just deploying it, are going to be the ones that pull ahead.
Fifth prediction: Cyber insurance starts influencing automation design directly.
If insurers decide your automation introduces too much operational risk, they’ll treat it the same way they treat weak MFA or bad backup hygiene — as an uninsurable condition. And that’s where personal accountability becomes real.
What insurers will look for is evidence that automated decisions are governed: critical actions have a human approval path, rollback procedures actually exist and are tested, and there’s a true kill switch for the automation stack. Providers who can show that become lower-risk. Providers who can’t will be fighting for coverage or paying for exclusions.
Sixth prediction: Shadow automation becomes a bigger problem than shadow IT ever was.
We already worry about users spinning up tools outside IT. Now add automation to that.
People are chaining together:
- AI tools
- Workflow platforms
- Financial systems
- Customer data
And they’re doing it fast, often without security review or legal oversight. By 2026, I think a lot of real business risk comes from automations IT didn’t build and doesn’t even know about. For example, imagine a marketing director chaining together a GenAI tool with a workflow platform that automatically approves invoices under a certain threshold and pushes them into the financial system. IT never sees it — but that one chain could trigger a fraud incident or a major compliance failure. That’s the shadow-automation risk MSPs will be expected to catch.
Seventh prediction: By 2026, “AI strategy” sounds late.
In 2024 and 2025, AI readiness made sense as a conversation.
In 2026, I think the question shifts.
It won’t be:
“How do we use AI?”
It will be:
“What shouldn’t be automated?”
AI becomes part of the plumbing. Strategy moves up a level to process design, controls, and outcomes. And that’s why the premium no longer lives in infrastructure management.
The providers who hold their margin will be the ones who reframe their work around risk — not just technical performance. They’ll be talking about the financial and regulatory consequences of automated decisions, about the governance needed to control them, and about the outcomes those systems are meant to deliver. That’s the advisory role customers are going to value, and it’s where the differentiation really lives as AI becomes background infrastructure.
Now, the predictions that might happen — but if they do, they move the whole industry faster.
Next prediction: The first automation-driven bankruptcy will get blamed on a “responsible AI failure.”
A well-intentioned, compliant automation that goes wrong at scale, which isn’t ransomware or a breach. When that happens, I think it forces a very real reckoning about how much trust we put into fully automated decision paths.
Another bold one: At least one MSP loses cyber insurance specifically because of unsafe automation design.
Not because they skipped MFA.
Not because they ignored backups.
Because the automation logic itself becomes the uninsurable risk. And once that policy language shows up in the market, everybody pays attention.
Final prediction: MSP valuation starts to split more cleanly based on automation maturity.
Two providers might have the same revenue and EBITDA, but very different multiples, because one is scaling through headcount and the other through governed automation. Investors are already watching efficiency metrics like Managed Endpoints Per Technician. If your MEPT is low, your multiple stays tight. If it’s high — driven by automation rather than people — your multiple expands. In 2026, that gap becomes visible.
So why do we care?
None of this is really about AI hype.
It’s about automation becoming the production system of IT services, identity becoming the control layer, and human responsibility never going away.
If you don’t control all three, you don’t really control the risk of the business you’re building.
If you’re not a bit comfortable right now remember: if nobody disagrees with you, you’re probably just describing the present.
And those are my 2026 predictions.