AI Automation Engineer

Insight Global; United States; On-site; USD 110,000 - 135,000; Full time

Job summary

Monument is seeking an AI Automation Engineer in the United States to build automation that replaces manual steps in operations, finance, HR, and sales support. You will deploy using Microsoft Copilot Studio, Power Automate, and Power Apps, integrating with Dataverse or SharePoint, and you will monitor live usage and iterate after launch.

The role emphasizes delivering usable automation, with code when needed (C#, Python, JavaScript) and strong collaboration with non‑technical stakeholders.

Qualifications

• 3-5 years building business applications, integrations, or automations that real people use daily.
• Hands-on delivery experience with Microsoft Power Platform, covering Power Automate and Power Apps at minimum.
• Working knowledge of Copilot Studio or a comparable conversational agent platform.
• Able to write real code when the platform runs out of road: C#, Python, or JavaScript, plus REST API integration.
• Demonstrated ability to map a business process and identify what is worth automating.
• Comfortable working directly with non-technical stakeholders and translating what they say into what they need.
• Clear written communication: documentation, runbooks, and training material people can use.

Responsibilities

• Build conversational agents in Copilot Studio: topics, knowledge sources, actions, and escalation paths.
• Build automations in Power Automate and applications in Power Apps, with Dataverse or SharePoint as the data layer.
• Integrate with the systems our teams already use, including Microsoft 365, Teams, SharePoint and internal line‑of‑business tools, using connectors, APIs, or custom connectors where needed.
• Write the code the low‑code platform cannot handle: custom connectors, Azure Functions, scripts, and API integration work.
• Apply the patterns, naming conventions, and environment standards set by the AI Solution Architect.
• Test against real cases including the ugly ones: malformed inputs, exceptions, and the edge cases the process owner forgot to mention

Skills

Power Automate

Power Apps

Copilot Studio

C# / Python / JavaScript

Process mapping

Stakeholder comms

Job description

Overview

Position Title: AI Automation Engineer

Position Summary:

We are looking for an AI Automation Engineer to build the agents and automations that Monument's own teams run on. You will sit with the people who do the work in operations, finance, HR, and sales support, watch how the work actually gets done, and replace the manual parts with something that runs itself.

This is an internal role. You are not billed to a client, and you do not carry an external account. Your stakeholders are the function owners inside Monument, and they are demanding in a different way than a client is: they sit next to you, they will tell you immediately when something does not work, and they have to live with what you build every day.

You build primarily on Microsoft: Copilot Studio, Power Automate, Power Apps, Dataverse, and the Microsoft 365 estate our teams already work inside. That is the default because it is where our data and our users already are, not because it is the only tool you are allowed to touch. When a problem is better solved another way, say so.

The measure of this job is use, not delivery. An agent nobody opens is a failure even if it works. You are expected to watch what happens after launch, fix what people actually trip on, and retire the things that did not earn their place.

How the Work Runs

Most of what you build will follow the same four steps. The first one is the one people skip, and it is the one that decides whether the rest was worth doing.

Steps:

• Sit with it: Watch the process being done by the person who does it. Count the clicks, the copy-pastes, the waiting, the rework. Do not work from a written process doc alone.
• Map and decide: Write down the process as it truly runs, including the exceptions. Then decide what to automate, what to simplify first, and what to stop doing altogether.
• Build the thin slice: Ship the smallest working version to real users in days, not months. Get it into the tools they already have open, such as Teams, Outlook and SharePoint, rather than a new place to go.
• Instrument and hand over: Add logging and usage tracking. Write the runbook. Train the owner. Make sure it survives without you in the room.

Responsibilities

Process Discovery

• Shadow internal teams to understand how a process genuinely runs, including the workarounds nobody documented
• Produce a clear current-state map: steps, handoffs, systems touched, time spent, and where the errors happen
• Size the opportunity honestly, in hours returned, errors avoided, or cycle time cut, before anything gets built
• Push back when the answer is not automation: some processes should be simplified, consolidated, or deleted first

Build & Configure

• Build conversational agents in Copilot Studio: topics, knowledge sources, actions, and escalation paths
• Build automations in Power Automate and applications in Power Apps, with Dataverse or SharePoint as the data layer
• Integrate with the systems our teams already use, including Microsoft 365, Teams, SharePoint and internal line‑of‑business tools, using connectors, APIs, or custom connectors where needed
• Write the code the low‑code platform cannot handle: custom connectors, Azure Functions, scripts, and API integration work
• Apply the patterns, naming conventions, and environment standards set by the AI Solution Architect
• Test against real cases including the ugly ones: malformed inputs, exceptions, and the edge cases the process owner forgot to mention

Adoption & Support

• Train the people who will use what you built, and write documentation they can actually follow
• Track usage after launch and act on what it tells you: where people abandon the flow, what they ask the agent that it cannot answer
• Own the fixes on what you have shipped. You are not done at handover
• Report on what each automation returned, in hours or errors, not in features delivered

Standards & Reuse

• Build to be reused: components, connectors, and prompt patterns that the next automation can pick up
• Contribute what you learn back to the internal pattern library so the same problem is not solved twice
• Flag governance and data‑handling concerns early: who can see what, and what should never reach a model
• Help evaluate new platform capabilities as Microsoft ships them, and say plainly whether they are worth adopting

The Stack:

Microsoft‑first, because that is where our users and our data already live. Depth in the agent and automation layers matters most; the rest you can learn here.

Agent: Microsoft Copilot Studio: topics, generative answers, actions, knowledge sources, agent flows

Automation: Power Automate (cloud and desktop flows), Power Apps, Power Platform environments and ALM

Data: Dataverse, SharePoint, Microsoft Graph, SQL, and the connectors between them

Intelligence: Azure AI Foundry, Azure OpenAI, prompt design, retrieval over internal content, evaluation of outputs

Code: C# or Python, REST APIs, custom connectors, Azure Functions, source control and deployment

Governance: Power Platform admin, DLP policies, environment strategy, Microsoft Purview, role‑based access

Qualifications

Required

• 3-5 years building business applications, integrations, or automations that real people use daily
• Hands‑on delivery experience with Microsoft Power Platform, covering Power Automate and Power Apps at minimum
• Working knowledge of Copilot Studio or a comparable conversational agent platform
• Able to write real code when the platform runs out of road: C#, Python, or JavaScript, plus REST API integration
• Demonstrated ability to map a business process and identify what is worth automating
• Comfortable working directly with non‑technical stakeholders and translating what they say into what they need
• Clear written communication: documentation, runbooks, and training material people can use

Strongly Preferred

• Microsoft certifications: PL-200, PL-400, PL-600, or AI-102
• Experience with Azure AI Foundry, Azure OpenAI, or retrieval‑augmented generation over internal content
• Dataverse data modelling and Power Platform ALM across environments
• Robotic process automation background: Power Automate Desktop, UiPath, or Automation Anywhere
• Experience automating back‑office functions such as HR, finance, recruiting, or operations
• Exposure to Power Platform governance: DLP policies, environment strategy, and admin tooling
• Prior work inside a professional services or consulting firm

Core Competencies

• Process Curiosity
• Build Velocity
• Platform Depth
• Stakeholder Partnership
• Adoption Focus
• Pragmatism
• Documentation Discipline
• Delivery Accountability

Pay Range

USD $110,000.00 - USD $135,000.00 /Yr.