The Aegentics blog
Field notes on
the agentic MSP
Notes on autonomous IT operations, service delivery, and the governance that lets ambitious MSPs move with confidence.
LATEST DEEP DIVE
The thinking behind autonomous service delivery.
Field notes
All field notes

AI Agents for MSPs: What They Do, How They Work and How to Choose One
AI agents for MSPs read tickets, act through your PSA, RMM and Microsoft 365 tooling, verify the result and document it, escalating what they cannot finish. This guide explains how they differ from copilots and scripts, which work they handle today, what your stack needs, how vendors price them, and the questions to ask before you buy.

What your L2 engineers do when L1 disappears
The easy tickets are going. Here is what is left.

Fixed price, variable cost: the half-finished industrialisation of managed services
MSPs industrialised how they sell but not how they deliver. Fixed monthly contracts sit on top of a cost base that moves with ticket volume. That is why a single vendor outage can turn a month negative.

In the 4th and current Industrial Revolution, AI disrupts IT operations fast and MSPs must respond faster
A word of comfort to warm up, IT managed services are here to stay and grow. AI introduces a host of new business challenges that increase demand for IT managed services, according to analyst predictions. But MSPs cannot satisfy this demand the same way. Managed services must, can, and indeed will become industrialized. The MSP industry that automated everyone never automated itself

The Full AI Governance Checklist for MSPs: Set Your Own Requirements and Score Any Vendor
Over the past six posts we have walked through the five critical control areas every MSP must examine before letting an AI agent touch a customer environment. The message from regulators on three continents is consistent: oversight must be structurally enforced, logs must be tamper-evident, personal data must stay out of the model, technical boundaries must be deterministic, and decisions that affect people must be explainable. The laws differ. The mechanisms they demand do not.

AI Transparency for MSPs: Labeling Output and Explaining Decisions Under Global Rules
Disclosure is the one AI obligation that is already binding almost everywhere. Korea fines companies that fail to notify users if they are interacting with AI. Utah requires disclosure in consumer interactions. Colorado will require pre-use notices and, after an adverse outcome, an explanation within 30 days. The EU adds a formal right to explanation for decisions that affect individuals.

Technical Guardrails for AI Agents: Why Prompts Alone Cannot Protect MSP Environments
What happens when the model is wrong, manipulated, or – what is called - hallucinating? That is the core architectural question for any agentic system. FINRA has already named the failure modes that matter most to regulated environments: autonomy without validation, scope creep beyond intended authority, and un-auditable decisions. A platform whose only defense is a carefully written prompt inherits all three risks.

Data Protection for AI Agents: Why the Model Must Never See Real Customer PII
Your tickets contain names, email addresses, phone numbers, national ID numbers, and free-text descriptions of your customers’ employees’ problems. The moment an AI platform sends that data to a third-party large language model, you inherit a cascade of obligations: a sub-processor question under GDPR, a business-associate question under HIPAA, and cross-border transfer restrictions across most of APAC and a growing number of US state privacy laws.

Tamper-Evident Audit Logs for AI Agents: The MSP’s Strongest Defense Against Liability
When an AI agent resets a password, provisions a user, or changes access rights, the action itself is only half the story. The other half is the evidence: who authorized it, what data the model saw, what decision path it followed, and whether that record can be trusted months or years later.

Human Oversight for AI Agents: Why Most “Human-in-the-Loop” Claims Won’t Survive an MSP Audit
AI agents promise massive and very real efficiency gains for MSPs; autonomously handling critical actions such as password resets, user provisioning, and access changes. But when that agent makes a decision that affects a customer’s system or employee, regulators across the globe have one clear message: human oversight must be real, enforceable, and documented.

Why MSPs Need an AI Governance Checklist Now
AI Governance for MSPs - The EU AI Act is in force, Texas TRAIGA (effective Jan 2026) offers safe-harbor through NIST alignment, South Korea’s AI Basic Act adds extraterritorial reach, and frameworks like Singapore’s agentic AI guidance emphasize the same core demands: enforced human oversight, tamper-evident logging, strong data protection, technical boundaries, and transparency. If governance is holding you back, you need an AI governance checklist to move ahead
