With Fulcrum MCP connector and AI Toolkit, workflows once too hard to build are now within reach

Every field organization has a list of workflows it decided not to build. Yours might have included confined space entry permitting, or a soil boring log that has to draw the layers, or a weld map that tracks non-destructive testing and coating status joint by joint down a pipeline run. Each one got looked at, considered, and set aside for a reason that made sense at the time.  

Fulcrum‘s Model Context Protocol connector, or MCP, and the AI Toolkit change what it takes to build those workflows. Both are available now through Fulcrum Labs, our early release program. You describe the process the way you would explain it to a new hire, and an end-to-end workflow comes back on the devices your crews already carry, joining the AI capabilities already running in the field.

Why field applications take longer to build than office applications

Building for the field means solving problems that office software never encounters, and none of them have anything to do with how complicated your process is. Your crews work past the last cell tower, so the app has to run offline and reconcile itself when coverage returns. A clearance measurement proves nothing without a location and a photo attached. Permissions have to hold across your staff and your contractors, with a name and a timestamp behind every edit.

Fulcrum settles all of that before anyone describes a single step of your process. Offline capture, spatial context on every record, media handling, permissions, the audit trail, and validation come with the platform, so a new workflow inherits them the moment it exists. None of it has needed describing for years.

What stayed hard was everything past the foundation. A workflow with real conditional logic, custom automation, or a report shaped to one discipline still needed someone who could build it — and even drag-and-drop takes time — so it waited in a queue behind whichever process mattered most.

What the MCP connector and AI Toolkit change

The two work as a pair. The connector lets an AI client build inside Fulcrum, and the toolkit teaches it how Fulcrum expects a workflow to be built. What they change depends on what you were trying to build.

• Simple workflows arrive in minutes. A form you would have assembled by hand comes back described and working, and the time it used to take stops being a consideration at all.
• Hard workflows stop waiting on a developer. The person who owns the process doesn’t have to know the words “conditional logic” to build it into a workflow. A custom report can appear the same day they decide they need one.
• The builds nobody thought possible become possible. A workflow with a drawing, a safety gate, or a strict sequence at its center still wants a technical hand. What used to take weeks now takes a few hours or days. 

Why nothing new lands on IT’s plate

A workflow built through the MCP connector runs on the same security model, permissions, and data structure as everything else you already have in Fulcrum. Nothing new shows up on your platform team‘s to-do list to secure, style, integrate, and then maintain. We have written before about what separates a vibe-coded field app from one that survives contact with a crew, and the difference comes down to what sits underneath it.

The MCP connector and AI Toolkit also work through the AI clients your organization already runs so nobody has to walk a new vendor through procurement. 

Which leaves the division of labor you want:

• People closest to the work do the building
• IT keeps security, compliance, and integration, without sitting in the middle of every request.

How App Extensions and Data Events led to the AI Toolkit

App Extensions and Data Events have been part of the platform for years. Technical teams have used both to build custom interfaces and automated behavior for work that no standard form covers. Fulcrum was built to be extended years before AI could write the extension.

Every extension built that way still required someone who could write one. The MCP connector and the AI Toolkit take that requirement away, opening the same extensibility to the people who understand the process without knowing how to code it.

Who can build today, and what comes next

Any customer ready to build can request access to Fulcrum Labs, and some already have. Their workflows are getting built now, and what they run into shapes what we fix next. The MCP connector and AI Toolkit work today, and they will be better in a month because of the teams using them.

Learn more: See the platform capabilities every workflow inherits.

The list is worth another look

Your crews have been covering for what never got built, with workarounds, retyping, and the steps that live in somebody‘s head. The people who understood those processes best had no way to fix any of it.

Fulcrum handles what every field job shares, so your crews can build the rest themselves.

Start with the one that has bothered you longest. Request access to Fulcrum Labs to build it, or join us on September 24 at 2 PM ET to see how to build field workflows you‘d given up on.