The job
Job Title: Automation and AI Developer
Reporting to: Programme Manager AI
Location(s): Hybrid, Manchester, Derby or Mansfield
About Us
Everything Tech Group is one of the UK's fastest-growing technology businesses, bringing together specialist expertise across IT, cloud, cybersecurity and communications. We're a growing technology business with ambitious plans, a collaborative culture, and a commitment to helping our colleagues succeed. By bringing together some of the industry's best talent, we've built a business where innovation, development and teamwork are at the heart of everything we do.
Job Purpose
The Automation and AI Developer designs, builds and maintains the automations and AI agents that ETG uses to deliver its managed services. The role works across the full tool stack: Rewst for workflow orchestration, Autotask as the PSA and system of record, ControlUp and Liongard for device and configuration telemetry, Microsoft Copilot Studio and Microsoft Foundry (Azure AI Foundry) for agents, GitHub for source control and release, and the security tools (Huntress, ThreatLocker, Mimecast and Inforcer) for detection and response.
The outcomes the role is accountable for are automated triage and closure of monitoring alerts, automated resolution of routine service requests, Copilot agents in Microsoft Teams for end users, engineers and account teams, and security response workflows that run with the right approvals and a complete audit trail. The role reports to the Programme Manager – AI and works day to day with service desk Technical Team Leaders, the build team and the Security Pillar.
Responsibilities
Workflow and integration development
- Design, build and test Rewst workflows that take triggers from ControlUp, Liongard, the security tools and Autotask and carry out the resulting actions, with the approval class, logging and rollback built into each workflow.
- Build and maintain the integrations between the stack components using REST APIs, webhooks and Microsoft Graph, covering Autotask, Microsoft 365, Entra ID, Intune and vendor APIs such as Veeam.
- Write supporting code in PowerShell, Python or JavaScript where a workflow needs logic beyond the platform's native capability.
- Maintain a library of reusable workflow components, forms and templates so that the same workflow can serve an agent, a form, an engineer and a scheduled job without being written more than once.
AI agent development
- Build and maintain Copilot agents in Microsoft Copilot Studio for end users, engineers and account teams. Agents call Rewst workflows rather than acting on customer systems directly, so that every action is approved, logged and reversible in one place.
- Build custom agents in Microsoft Foundry (Azure AI Foundry) where a use case needs more than Copilot Studio provides: model selection, custom tools and function calling, retrieval over ETG and customer knowledge, evaluation and monitoring.
- Design topics, prompts, knowledge sources and guardrails; test agents against real ticket data; measure accuracy, resolution rate and escalation rate before and after release.
- Build and tune the ticket classification used to route alerts and human requests into automatable categories, working from Autotask ticket data.
Service desk automation
- Automate the triage and closure of monitoring alerts, including backup notifications, infrastructure and RMM alerts, expiry reminders and network alerts, with human sign-off retained where required, in particular for security and identity alerts.
- Automate routine service requests such as password resets, joiner, mover and leaver processes, mailbox and licence changes and standard change requests.
- Work with Technical Team Leaders and service desk engineers to validate each automation against live tickets before it goes into production, and hand it into business-as-usual operation with a runbook and a named owner.
Security workflow automation
- Build response and containment workflows for Huntress, ThreatLocker, Mimecast and Inforcer detections, using the agreed trigger confidence tiers, with Service Desk sign-off for containment actions.
- Build configuration drift and compliance workflows from Liongard and Inforcer baselines, including Microsoft 365 tenant settings, conditional access and admin role changes.
Quality, governance and documentation
- Apply the action classification model to every workflow: automatic, confirm before acting, or human only. Keep an audit trail for every automated action.
- Keep workflow definitions, agent code, prompts and configuration in GitHub, with pull request review and GitHub Actions for testing and release. Test in a non-production tenant, raise change records and keep documentation and runbooks current for every workflow and agent.
- Monitor workflow health, failures and run costs; fix issues promptly; report monthly on automation volumes and outcomes to the Programme Manager – AI.
- Follow ETG's information security policies and customer contract terms in the handling of customer data, including least-privilege service accounts and secrets management.
Measure of Success
4. Measures of success
Performance will be assessed against the following measures, with targets agreed with the Programme Manager – AI and reviewed quarterly.
MeasureWhat it coversAlert automationMonitoring alerts auto-triaged or auto-closed per month, and the share of total alert volume handled without engineer involvementRequest automationRoutine service requests resolved by workflow or agent per month, and the share of request volume automatedWorkflow reliabilityFailure rate, time to fix and number of rollback incidents across production workflowsAgent performanceResolution rate, escalation rate and accuracy of Copilot agent interactions, measured against test sets and live useEngineer time releasedEngineer hours released by automation, measured against the baseline agreed at the start of each build phaseDelivery against planWorkflows and agents delivered against the agreed build sequence and datesDocumentation and handoverEvery production workflow has a runbook, a named owner, a test record and a change record5. Experience Qualifications, Personal Attributes
AreaEssentialDesirableExperienceAt least three years building production automations and integrations in a managed services, IT services or IT operations environment.
Hands-on experience with an automation or orchestration platform such as Rewst, Power Automate, Azure Logic Apps or n8n, and with building integrations against REST APIs and webhooks.
Experience with the Microsoft 365 platform, including Microsoft Graph, Entra ID and Intune.
Experience building AI agents in Microsoft Copilot Studio and custom agent builds in Microsoft Foundry (Azure AI Foundry), Azure OpenAI or a comparable framework, including tool and function calling, retrieval and evaluation.
Day-to-day use of GitHub for source control, pull requests and GitHub Actions.
Direct experience with Rewst, Autotask, ControlUp, Liongard, Huntress, ThreatLocker, Mimecast or Inforcer.
Experience working with PSA ticket data and classifying tickets for automation.
Experience in an MSP service desk or NOC, so that alert noise and engineer workflow are understood first hand.
Experience with Veeam, Microsoft Purview or the wider Power Platform (Power Apps, Dataverse).
Skills and knowledgeScripting in PowerShell and at least one of Python or JavaScript.
Working knowledge of JSON, OAuth 2.0, API authentication, rate limits and error handling.
Understanding of incident, request and change processes in a managed services environment.
Able to design a workflow with approval steps, logging and rollback, and explain it clearly to a non-technical colleague.
Microsoft certification such as PL-400 (Power Platform Developer), AI-102 (Azure AI Engineer) or MS-102 (Microsoft 365 Administrator).
Experience with the Microsoft Agent Framework or Semantic Kernel, and with GitHub Copilot as a development tool.
Knowledge of data protection requirements when customer data passes through automated workflows.
Personal qualitiesPractical and outcome-focused: delivers working automations rather than prototypes.
Methodical about testing, documentation and change control.
Works well with service desk engineers and team leaders and builds their feedback into the work.
Learns new platforms quickly and keeps up with changes in the Microsoft and vendor toolsets.
Comfortable demonstrating work to internal stakeholders and customers.Qualifications and other requirementsRight to work in the UK.Degree in computing or a related subject, or equivalent experience.
Relevant Microsoft or vendor certification.
Key Relationships
Internal
- Technology and Engineering Teams