Ask a personal injury firm’s marketing director where last month’s signed cases came from, and you will get a spreadsheet. Local Services Ads, directories, paid search, organic, referrals: each with a cost per lead and, at the better firms, a cost per signed case.
Agent-sourced demand is not on it. When an agent researches firms and the person then calls one, the call arrives with no referrer. It lands in “direct”, or “organic”, or nowhere. The step that did the research gets no credit, and nobody can tell whether it is growing.
What a request carries
When the request is made on your site through your Agent Concierge, it carries its origin from the first moment:
- the agent client it came through, when the agent identifies itself;
- the surface: the tools on your page, your MCP endpoint, or an express page;
- the person’s own words, as they asked;
- what the agent gathered;
- the consent the person gave, per channel.
Later, it carries what happened. Your team can mark the outcome in the inbox. Once your texting registration is approved, a staff text can ask “New client? Reply Y or N.” And the CRM adapters that report outcomes back are built, each waiting on a sandbox run. An outcome can carry a value, so the value of a signed case is recorded with it.
From requests to a P&L
Put those together and you get a lead-source P&L: one table, a row per source, counting leads, confirmed requests, bookings and signed clients, with their value. The value is what your team reported for a signed client, or the estimate until they report one.
The agent rows are split by how the agent reached you:
- the tools on your page;
- an agent working from its own servers, through your hosted MCP endpoint or an express page;
- a tracked call.
A separate funnel breaks the same requests down by agent client.
Set it next to the sources you already report, and it starts to answer the question the spreadsheet cannot: what demand that arrives through an agent is worth to you. You can compare it with what you pay the directories.
It also answers a quieter question: which agents send you requests that get booked, and which send you volume. The answer shapes what you publish and where you spend.
Feed the winners back
Knowing which requests became clients is also the signal an ad platform needs. Confirmed conversions can go back to the ad platforms where the rules allow it, so the platforms learn which clicks become clients, not which clicks fill in forms.
The rules differ by vertical, and they are in the code:
- elective health: conversion uploads to Google Ads only, and no audiences;
- housing: conversion uploads, and no audiences.
Why it matters now
Agent-sourced demand is small today for most businesses. That is exactly when you want to measure it. By the time it is large, the businesses that measured it early will know which agents send them clients, and what those clients were worth. The ones that did not will be looking at a growing “direct” line and guessing.
Where this stands
Trace, the lead pipeline, outcome reporting and the ad rails are built and deployed. We do not have a customer’s numbers to show you yet, and we will not invent any.
Want agent-sourced demand on your lead report? Run the free audit to see what an agent finds on your site today, or book a demo.