- Service calendar
- Scorecards
- Account crosswalks
- Realized and unrealized gains tables
Instead of a team, a consultant or a call to the vendor.
Connect your AI assistant (Claude, ChatGPT, Codex or Gemini) to your Milemarker platform. Query data, build dashboards and workflows, manage users and run projects in plain conversation, signed in as yourself and within your permissions. Two-minute setup. Zero middleware.
Request AccessI'll pull outflows from Snowflake and cross-reference recent meeting notes for context.
Top 5 outflows for Q1 2026:
Pattern: three of these are rollovers — likely retention opportunities. Want me to draft outreach notes, or pull the advisors of record?
Built on the open MCP standard. Your firm picks the AI. Milemarker powers the platform underneath.
HOW IT WORKS
Follow one request from your AI assistant into Milemarker and back. Your role, your entitlements and your approval come along at every step.
Plain language in the assistant you already use. No SQL, no report request.
The connection carries your Milemarker role. There’s no separate AI login with broader access.
Entitlements, row-level security and PII masking are enforced beneath the MCP, so they can’t be prompted away.
Reads run right away against your reconciled data, through the same governed APIs as the app.
A governed answer scoped to your book, with personal data masked per policy.
“Which of my clients have cash balances above their target range?”
Describe the outcome you want, not the syntax.
Building needs build-level access in your role. The MCP confirms which firm before it touches anything.
New metrics and pages inherit your entitlements and PII masking. Nothing to re-implement per build.
It drafts the build and returns a pre-build summary: what it will create, where, and who will see it.
You approve, it goes live in Milemarker, and you check it in the app.
“Build a dashboard of AUM by office for the leadership team.”
Nothing is created until you approve.
Admin work in a sentence instead of a day of admin screens.
Only people with admin rights can manage users, and the MCP can’t widen anyone’s access, including yours.
The change is scoped to your firm. Other firms’ data is never in reach.
It shows a preview of the change that doesn’t touch anything until you confirm.
Confirmed, the invitation goes out and the change is on record.
“Add the new advisor to the Denver team with read-only access.”
Review the change, then confirm.
You also decide, tool by tool, what your assistant may do:
HOW FAR IT GOES
Most AI connections stop at answering questions. The Milemarker MCP reaches every layer of your platform, from the feeds coming in to the workflows going out, with your governance at every step. It’s not a chatbot on your data.
Feeds, files and firm-defined records, on your schedule.
“I want my Black Diamond data to sync at 3 PM Eastern instead of 8 AM.”
One definition of a household, and who may see what.
“Create a non-Navigator access group and add these 12 users to it.”
Answers from reconciled data, saved as numbers everyone agrees on.
“Which households are underperforming their blended benchmark?”
What each advisor, office and client sees, on web and mobile.
“Based on how you understand this data, build the best possible filtering for me.”
Repeatable processes, tracked, pre-filled and connected outward.
“I need an address change form that links out to my custodian.”
Reads run instantly. Anything that changes your platform shows you a summary first and waits for your approval.
“Firm MCP brings all of it into one foundation that knows how you define a household, who is allowed to see what, and what your numbers mean. That foundation belongs to the firm. And it's live right now. Our clients are using all of it today, with no queue, and it's changing how their teams work this week.”
Jud MackrillCo-Founder & CEO, Milemarker 690+
Hours a year handed back to each MCP user
$100K+
A year in consultant time per user that firms don’t have to buy
⅓
Of a full-time role gained for every MCP user
22%
Of platform changes now made by firms’ own staff, and rising
Annual figures are a forecast: a third of a 2,080-hour work year per user, below the pace measured across client firms in early October 2026. Consultant time valued at $150 an hour, the low end of typical U.S. rates for experienced platform consultants. Share of changes measured in September 2026.
IN THE FIELD
These come from client teams, in their own words at our MCP office hours. No tickets, no consultants, no waiting on a vendor queue.
Instead of a team, a consultant or a call to the vendor.
Runs the platform their own way, entirely by chat.
Started as a static HTML mockup, then brought to life with live data.
Published inside the platform instead of behind another login.
“Every new AI tool used to mean another system to govern. Milemarker's Firm MCP gives Root one foundation across every system we run, with our permissions already in place. Our operations team can build what they need the same day, and compliance has one place to oversee it.”
“Custodians, CRMs and meeting tools are all opening up through MCP, and that gives my teams real leverage. What makes it work at firm scale is having our own layer underneath, one place where our data and our definitions live together. That's a real moat for advisory firms. Milemarker Firm MCP is the layer we're building that on at RFG, so what we ship fits the way our advisors actually work.”
MILEMARKER MCP
MCP — Model Context Protocol — is an open standard. Think of it as USB for AI. Just as any USB device works with any computer, any MCP-compatible AI assistant can plug into any MCP-compatible system.
Milemarker is an MCP-compatible system for wealth management, so your firm isn’t locked into one AI vendor. Connect Claude today, add Gemini tomorrow, or run both side by side. Governance lives in Milemarker, underneath the MCP, so every assistant you connect inherits the same guardrails.

THE GUARDRAILS DON'T GO AWAY
You connect with your own Milemarker account. The AI acts on your behalf, with your role, and never with broader access.
If you can’t see a household in the app, the AI can’t see it through the MCP. Tenant boundaries keep every firm’s data separate.
Masking and row-level security are applied at the data layer, before anything reaches the AI conversation.
Reads are instant. Every build or change shows a summary first and waits for your approval. The highest-impact actions also need an approval click in the Admin Console.
Each change produces a reviewable artifact, so your team can see exactly what was done and by which request.
The MCP runs inside Milemarker’s SOC 2 Type II environment. Milemarker is your data processor; your firm stays the data controller.
“Milemarker and its MCP help bring [our] ecosystem together, connecting data across CRM, reporting, planning and other systems into one consistent, permission-aware view. That means we can go beyond giving advisors a complete picture of their clients to surface what matters, explain why it matters and identify what to do next. That shift from connected data to actionable intelligence is where the real value lies.”
“I know what the data is. I know what it should be. I know where it lives and what number we're trying to get to. I used to write out the whole recipe, hand it to an engineer and wait for someone to get to my report. With the Milemarker MCP, I just do it myself, across every system we run, with our permissions already in place. It's been magic for me.”
GET STARTED
Milemarker turns on Firm MCP for your firm. Add your firm’s MCP address as a connector in your AI client, then sign in with your own Milemarker account. MCP is live now and rolling out across Milemarker partner firms.
Questions? connect@milemarker.co
Whether you're exploring AI interfaces for the first time or evaluating how MCP fits your firm's compliance requirements, find what you need below. Reach out to our team at connect@milemarker.co
Milemarker MCP is a server that exposes your entire Milemarker platform — data queries, dashboard deployment, user management, search, and AI navigators — as something any MCP-compatible AI assistant can operate. Instead of clicking through admin screens or writing SQL, your team asks in plain language and the platform does the work.
Any AI assistant that supports the Model Context Protocol standard. Today that includes Claude (Anthropic), ChatGPT (OpenAI), Codex (OpenAI), and Gemini (Google) — plus any future model that adopts the open standard. Your firm is never locked in.
A chatbot answers questions about information it’s been trained on. Milemarker MCP is an operator — it takes action on your live platform. When you say “deploy a dashboard,” it actually builds and deploys the dashboard. When you say “add a user,” the user is created with real permissions. Every action goes through Milemarker’s governed APIs with a full audit trail.
Yes. Every action flows through Milemarker’s existing governed APIs, respects tenant boundaries, and produces auditable artifacts. Nothing skips policy. Nothing happens in a black box. Milemarker is SOC 2 Type II certified with 256-bit SSL encryption and 24/7 threat monitoring.
About two minutes. Milemarker turns on Firm MCP for your firm, you add your firm’s MCP address as a connector in your AI client, and you sign in with your own Milemarker account. No engineering resources required.
Yes — Milemarker MCP operates on top of your Milemarker platform. It’s the AI interface layer for the data, dashboards, users, and workflows already running in your Milemarker tenant.
Query your firm’s data in natural language, deploy dashboards and metrics, stand up search and AI navigators, manage users and permissions, and run platform-wide projects like bulk deployments, migrations, and audits — all through conversation.
CONNECT EVERY SYSTEM IN YOUR STACK TO A SINGLE, INTELLIGENT WORKFLOW ENGINE—BUILT ON YOUR UNIFIED DATA AND READY TO SCALE.
We use cookies to enhance your experience, analyze site traffic and deliver personalized content.
Read our Cookie Policy