The Shift Is Already Here
In 2025, one of Australia's largest MSPs quietly reduced its Level 1 helpdesk team by 40% β not through layoffs, but through attrition. They simply didn't backfill roles because an AI triage system was handling first-line tickets.
In 2026, that same MSP is rolling out AI-generated documentation, automated patch management, and AI-driven network monitoring that flags anomalies before they become incidents.
This is not a prediction. This is happening now.
Here are the six roles that will be transformed or eliminated in MSPs within the next 18-24 months β and what to do about it.
1. Level 1 Helpdesk Engineer
Why It's Disappearing
AI triage systems can now: - Classify and prioritise tickets by severity - Answer common questions from a knowledge base - Run basic diagnostic scripts automatically - Escalate only the genuinely complex issues to humans
The math is brutal. An AI triage system costs a fraction of an L1 salary and works 24/7. It doesn't call in sick. It doesn't burn out.
What's Left
L1 roles won't vanish entirely, but they will shrink dramatically. The remaining positions will require technical depth β not just script-following. If your L1 role consists mainly of password resets, mailbox creation, and ticket triage, your timeline is 12-18 months.
The Escape Route
Move to L2 or specialise. Learn automation β if you can build the AI tools instead of being replaced by them, you've flipped the dynamic.
2. Report Generator / Data Entry
Why It's Disappearing
AI can generate client reports, compliance documentation, and monthly service reviews in seconds. Models like GPT-4 and Claude can ingest raw monitoring data and produce polished PDF reports with zero human intervention.
What's Left
Strategic analysis β interpreting what the data means, not just formatting it. Engineers who can tell a client "your security posture improved because of X, but you're vulnerable to Y" will be valued. Engineers who just paste numbers into a template won't.
3. Basic Network Monitor
Why It's Disappearing
AI-driven monitoring platforms (Cisco AI, LogicMonitor, Datadog) now detect anomalies, predict failures, and even auto-remediate common issues. The human-in-the-loop is increasingly optional for routine monitoring.
What's Left
Complex incident response and architecture design. The human value shifts from "watching dashboards" to "designing the monitoring system" and "handling the 5% of incidents the AI can't solve."
4. Manual Scripting / Automation Engineer (Basic)
Why It's Disappearing
Ironically, the AI that's replacing other roles is also replacing the people who write basic automation scripts. Copilot, Claude, and ChatGPT can generate PowerShell, Python, and Bash scripts from natural language prompts.
An engineer who previously spent 2 hours writing a migration script can now do it in 10 minutes with AI assistance.
What's Left
Automation architecture β designing the system, not writing the code. The premium shifts from "can you write a PowerShell script?" to "can you design an automated workflow that spans RMM, PSA, and client environments?"
5. Password Reset / Account Admin
Why It's Disappearing
Self-service password reset (SSPR), automated account provisioning via Entra ID, and AI-driven identity governance have made this role largely obsolete. Modern MSPs set it up once and it runs itself.
What's Left
Identity architecture and security policy design. The people who design the Entra ID tenant, configure conditional access policies, and manage privileged identity management will remain. The people who manually create accounts won't.
6. Basic Documentation Writer
Why It's Disappearing
AI can generate documentation from ticket histories, change logs, and system configurations. Tools like Notion AI, Guru, and custom LLM pipelines now produce client-facing documentation that's more consistent and up-to-date than manually written docs.
What's Left
Knowledge management architecture β designing the systems and workflows that feed the AI. The skill shifts from "writing docs" to "designing how knowledge flows through the organisation."
What Engineers Must Do Now
The MSP industry is not dying, but the role mix is shifting dramatically. Here's your survival plan:
Short-Term (0-6 months)
- Automate your own job. If you can describe a task in a prompt, it can be automated. Do it before your employer does.
- Learn AI tooling. Copilot, ChatGPT, Claude β become proficient. The engineers who use AI effectively will replace those who don't.
- Document everything you do. Not for job security β for the portfolio that gets you your next role.
Medium-Term (6-18 months)
- Specialise in security. This is the highest-growth, most AI-resistant area in IT. Essential 8, SOC, threat hunting, DFIR.
- Move up the stack. From break-fix to architecture. From following runbooks to designing them.
- Develop client skills. Relationship management, consulting, pre-sales β these are hard to automate.
Long-Term (18+ months)
- Become the person who designs the AI systems, not the person they replace.
- Build a personal brand. Write, speak, share. The engineers with visible expertise will always have options.
- Stay portable. Don't let golden handcuffs (training bonds, non-competes) trap you in a dying role.
The Verdict
AI is going to reshape the MSP workforce. Some roles will shrink. Some will disappear. But the demand for skilled engineers who understand both technology and business is only going up.
The question isn't "will AI take my job?" β the question is "am I positioning myself for the roles AI creates, or clinging to the roles it eliminates?"
Related Reading
- AI in the Service Desk: Job Killer or Force Multiplier?
- MSP AI Automation Impact 2026
- MSP Engineer Career Paths: From L1 to Architect
- MSP Burnout Survivors: Engineers Who Quit and Thrived
- Best Certs for MSP Engineers 2026
- Remote Monitoring and Management (RMM) Guide
Was this helpful?