Three people need the same thing at different moments. A partner overseeing a portfolio of clients wants to know which engagements have overdue requests before a client call, without clicking into each one. A manager prepping a status meeting needs every open item across their active engagements in a single view. A preparer starting the day wants to know where to focus first instead of checking each engagement on its own. All of it lives in Fieldguide, and all three answers are usually somewhere else when the question comes up: a client email thread, a Slack channel, a document.
So the routine kicks in. Open the platform, click into the engagement, scan the request list, copy the relevant pieces back into the conversation. Or skip all that and ping the senior, who answers the same question for the third time this week. Firms have been asking us a direct question for a while now: can the AI tools our teams use every day connect to Fieldguide? Starting this month, they can. Fieldguide MCP is rolling out to firm customers, and it puts live engagement data inside the generative AI tool your firm has already standardized on. For the partner, portfolio status becomes a question typed where they already work. For the manager, the pre-meeting rollup assembles itself. For the preparer, the day starts with priorities instead of tabs.
Engagement answers inside the conversation you're already having
Fieldguide MCP is built on the Model Context Protocol, the open standard that lets AI tools connect to outside systems. In practice, that means Fieldguide sits alongside connectors for the tools teams already work in (like Slack, Google Drive, and the Microsoft suite, for e.g.), in your team's AI workspace, and a plain-language question gets a real answer. "What's the status of the Medline SOC 2 engagement?" returns the period, report date, open and overdue requests, and owners in a single response. "Show me everything open across my active engagements" returns exactly that, with overdue items flagged.
Because the connection lives next to your other tools, the engagement record becomes part of the same conversation where the rest of the work is happening. A manager working through a client thread can pull engagement status into that conversation and respond without leaving it. The same data that powers the Agent Workforce is now reachable from the tools your firm uses every day.
Answers that match what's in Fieldguide right now
Every look-up runs against live Fieldguide data. There is no export, no sync, no snapshot from last night. If a client uploaded evidence ten minutes ago and the request closed, the answer reflects it.
Access works the same way. A user can only look up or act on data they already have access to in Fieldguide. Connecting through Claude or Copilot grants nothing extra; the permission model in the platform is the permission model in the conversation. If you lead or own engagements, questions about "my engagements" filter to yours automatically, and you can always name the specific engagements you want.
Routine actions, approved before anything changes
Look-ups are half of it. This release also supports routine actions you can take without leaving the conversation: post a follow-up comment on overdue requests or start a new engagement from a standard template.
It’s important to note that nothing changes without a yes. Before any action runs, the AI tool shows exactly what it plans to do, down to the text of the reminder comment, and waits for approval. Look-ups happen directly; changes always ask first.
Connected without a custom build
Connecting Fieldguide takes a few minutes and no engineering effort. Generate an access token in Fieldguide account settings, add Fieldguide in your AI tool's connector settings, and paste the token. That's the whole flow, and it is the same in any MCP-compatible tool.
Because Fieldguide MCP is built on an open standard, the same connection works whether your firm has standardized on Claude, Microsoft Copilot, or Google Gemini. Nothing changes on Fieldguide's side if you switch tools or add another. There is no integration to scope, no middleware to maintain, and no dependency on which AI vendor your firm picks next year.
For partners: Walk into every client call current
Getting up to speed before a client call has historically required partners to ask their engagement team for the latest or piece status together from email threads and various status meetings. Now the status brief is delegated to Fieldguide MCP: ask ten minutes before the call, in the same conversation where your calendar and meeting notes already live, and walk in knowing what's open, what's overdue, and who owns each item. When the client asks about a specific request mid-call, you can answer in the moment instead of promising a follow-up.
After the call, the discussion turns into action without a trip into the Fieldguide platform: ask for a reminder comment on anything overdue, review the exact wording, approve it, and draft the recap with the engagement status already in context. And when the client calls back later in the week, the answer reflects the engagement as it stands, not as it stood at last week's status meeting.
For managers: Monday morning without the assembly
Managers are expected to know where every engagement stands: what is open, what is overdue, what the client owes, and what the team should work on next. The week used to start with assembly, opening each engagement in turn to build that picture. Now the first read is one prompt across the whole portfolio, covering what moved over the weekend, what is stuck, and where the week's risk sits.
From there, the setup runs in the same conversation. Triage the overdue items by owner, ask for the reminder comments, review the wording, and approve; the nudges are out before the first meeting. See what's due this week, set the team's priorities from a current read, and draft the partner update without leaving the thread. And when new work kicks off, start the engagement from the firm's standard template. Monday morning goes to setting priorities rather than collecting them.
For staff: Start the day with priorities, not tabs
Working several engagements at once fills the day with small detours: what's due next, whether anything is blocking a request, what the client sent overnight. Each one used to mean setting aside the task in front of you to go look. Now the day opens with a prioritized worklist across every engagement, with anything overdue flagged, and the mid-task check happens in the AI tool where the memo is already being drafted. The check takes a moment and the task keeps moving.
The status questions from other people shrink too. When a manager or senior pings for status, the answer comes from live data in the same conversation. After a client upload lands, one question shows what came in and what's still outstanding, and at the end of the day the same ask surfaces tomorrow's follow-ups before you close the laptop.
Popular prompts to try first
The fastest way to understand Fieldguide MCP is to ask it something you already needed to know this morning. There is no query language to learn and no menu to memorize; the questions that work are the ones your team already asks each other. Here are a few that firms reach for early:
- Check status: "What's the current status of the [client] engagement?"
- See what's open: "Show me all open requests across my engagements."
- Take a routine action (approved first): "Post a reminder on the overdue requests for [engagement]."
- On our roadmap (Insights): "Which engagements are at risk of a missed deadline?"
The feature is meant to be explored, so ask for what you need in your own words, and every answer comes back from live engagement data, scoped to what you can already see.
Your starting point for AI-connected engagement work
Managers and partners who check status across a portfolio of clients feel this immediately: the standing "what's still open?" question stops requiring a trip into the platform or a message to the team. Preparers get a faster start to the day and fewer interruptions from status requests they used to field manually. And firms evaluating how Fieldguide fits into their AI environment get a concrete answer: the engagement record connects to the tools you already run, through an open standard, with your existing permissions intact.
Fieldguide MCP is rolling out to firm customers now. Reach out to your Fieldguide account team with any questions or for more information.

Nick Giantis
Principal Product Manager
Related posts
View All Posts
Fieldguide Launches Field Financials and Fieldguide MCP, Along With a Refreshed Brand
- News

Building the Agentic Firm: Announcing Field Financials, Fieldguide MCP and Our Refreshed Brand

Review Is the New Prepare: The Auditor's Job in an Agent-First World
- Solutions

Grant Thornton Advisors Partners with Fieldguide to Power Next-Generation Risk and Controls Services with Agentic AI
- News