← ALL BLUEPRINTS / AGENT BLUEPRINT
Revenue analyst agent for professional services firms on Salesforce with conflicting marketing and sales numbers
A custom AI data agent that connects Salesforce, the marketing platform, time tracking and finance through MCP connectors, applies one agreed set of revenue definitions, and answers leadership questions in plain English with the source shown. It replaces the quarterly argument between marketing and sales with a single weekly executive digest that both teams sign off on.
| ROLE | FOUNDER / LEADERSHIP | PLATFORM | CUSTOM |
|---|---|---|---|
| INDUSTRY | PROFESSIONAL SERVICES / AGENCY | ||
| DRAWN BY | HACKING DEMAND | ISSUED | OCTOBER 24, 2026 |
What The Revenue Analyst does
Leadership at a 51 to 200 person professional services firm sits through the same meeting every quarter: marketing says it sourced one number of pipeline, sales says another, and the founder cannot tell which one to plan hiring and partner capacity against. The gap usually comes from different definitions (what counts as a lead, which date a deal lands in a quarter, how referrals from partners and past clients are attributed) applied to different systems. The Revenue Analyst works for the founder and leadership team. It owns a written metric dictionary agreed by marketing, sales and finance, reads live data from Salesforce and the surrounding systems through MCP connectors, answers ad hoc questions with the exact definition and records used, flags when the two teams' numbers diverge and why, and produces a weekly executive digest covering pipeline, proposal throughput, win rate by source (including referrals), and partner utilization against booked work. It is built and maintained as a managed service, so definition changes, broken connectors and new questions are handled without the firm needing an in-house data team.
Custom build
The firm runs Salesforce, so HubSpot Breeze is not the right base, and Salesforce's native reporting is part of the problem because each team built its own reports with its own filters. A custom AI build fits better: MCP connectors let one agent read Salesforce, time tracking, proposals and finance together, which is what a services firm needs to tie pipeline to partner capacity. n8n handles the nightly syncs and the digest schedule, a small warehouse or Airtable base holds the reconciled model, and Databox gives leadership dashboards for drill-down. Keeping the metric dictionary outside any single team's tool is what stops the numbers from drifting apart again.
What starts it
- Monday morning schedule fires to assemble and send the weekly executive digest
- A leader asks a question in Slack or Teams, such as 'what did referrals contribute to closed revenue last quarter?'
- A reconciliation check finds marketing-reported and sales-reported figures for the same period differ beyond an agreed tolerance
- An opportunity sits in proposal or scoping stage longer than the agreed threshold
- Quarter close, when the agent prepares a locked quarter snapshot for the leadership review
What it reads
- Salesforce (opportunities, stages and stage history, close dates, opportunity source and referral fields, accounts, campaigns)
- Marketing automation platform and web analytics (form fills, campaign membership, first and last touch)
- Toggl or the firm's time tracking system (billable hours and utilization by partner and practice)
- Accounting or invoicing system (invoiced and recognized revenue per client engagement)
- Proposal tool such as PandaDoc (proposals sent, viewed, signed, and value)
- A shared metric dictionary stored in a versioned document or table that defines every reported number
How it works, step by step
- Agree the metric dictionary firstRun a working session with marketing, sales, finance and the founder to write one definition for each contested number: sourced pipeline, influenced pipeline, referral revenue, booked versus invoiced revenue, quarter assignment rule (created date versus close date), and win rate. Each definition names the Salesforce fields and filters it uses. The agent will only report against these definitions.
- Connect systems through MCP and a light model layerExpose Salesforce, time tracking, proposals and finance to the agent through read-only MCP connectors. For heavier joins (utilization against pipeline, source attribution across systems), sync nightly into a small warehouse or Airtable base using n8n, so queries are fast and repeatable rather than recomputed differently each time.
- Reconcile the marketing and sales numbersFor each period the agent computes both teams' versions alongside the agreed definition and produces a bridge: which opportunities are counted by one side and not the other, and why (missing source field, duplicate account, deal moved quarters, referral logged as a campaign). Record-level fixes go to a data hygiene queue for the Salesforce admin.
- Answer leadership questions with receiptsWhen a leader asks a question, the agent identifies the matching definition, runs the query, and replies with the number, the definition used, the record count, and a link to the underlying Salesforce report. If no definition covers the question, it says so and proposes one instead of improvising.
- Watch the proposal bottleneckTrack time from discovery to proposal sent, proposal to signature, and how many open proposals each partner is carrying against their utilization. Surface stuck scoping work and partners who are both overbooked on delivery and holding several unsent proposals.
- Build the weekly executive digestEvery Monday assemble a one page digest: pipeline and bookings versus target, win rate by source with referrals broken out by referrer type, proposals in flight and aging, forecast utilization by practice for the next eight weeks, and any reconciliation gaps opened or closed. Send to leadership by email and Slack, with Databox dashboards for drill-down.
- Lock the quarterAt quarter close, snapshot the agreed numbers so later edits in Salesforce do not change what leadership reviewed. Changes after lock are logged as restatements with a reason.
When a person takes over
- When a metric definition is ambiguous or a new question needs a new definition, the agent drafts options and the revenue operations owner (or founder) decides; the agent never invents a definition on the fly.
- When reconciliation finds records with missing or conflicting source data (for example a deal tagged both as a partner referral and a paid campaign), the Salesforce admin or account owner corrects the record, because only the person who ran the deal knows the true origin.
- When a number in the digest moves sharply week over week or a locked quarter needs restating, the finance lead reviews before it goes to the full leadership team.
- When a leader asks about an individual partner's performance or utilization for compensation or review purposes, the agent provides the data to the managing partner only and does not draw conclusions.
Rules it follows
- Read-only access to every system; the agent cannot edit Salesforce records, invoices or timesheets, only flag them.
- Every number it reports cites the dictionary definition, the date range and the record count; if data is incomplete it says how incomplete rather than estimating.
- Referral attribution follows the agreed rule (for example, a referral from a past client or partner outranks any later marketing touch) and the agent never reallocates referral credit to campaigns to make channel numbers look better.
- Partner-level utilization, billing rates and individual relationship data are restricted to named leadership roles and never appear in channel-wide posts.
- Booked, invoiced and recognized revenue are always labeled separately, since in services work a signed scope is not yet delivered revenue.
How you know it works
- Gap between marketing-reported and sales-reported pipeline for the same quarter, tracked to zero unexplained variance
- Share of leadership revenue questions answered by the agent without an analyst pulling a manual report
- Percentage of open opportunities with a valid source and referral field populated
- Median days from discovery call to proposal sent, and proposal to signature
- Weekly digest delivered on time and acknowledged by leadership each week
From pilot to live
- Pilot: definitions and one quarter (weeks 1 to 4)Write the metric dictionary, connect Salesforce and the marketing platform read-only, and reconcile last quarter's disputed numbers. Leadership reviews the bridge report and confirms the agent's figures against a manual check before anything is automated.
- Expand: digest and utilization (weeks 5 to 10)Add time tracking, proposals and finance connectors, start the weekly executive digest in parallel with existing reports, and open the Slack question channel to the founder and two or three leaders. Tune answers based on what they actually ask.
- Full rollout and ongoing maintenanceRetire duplicate team-level reports in favor of the digest, open questions to all practice and sales leads with role-based access, and move to a managed cadence: monthly definition review, connector monitoring, quarter-close locking, and new metrics added as the firm's questions change.
Tools that fit
n8n
Workflow automation platform
Databox
Dashboards and AI analytics
Airtable
Visual data organization
Toggl
Timetracking & productivity tool
PandaDoc
Digital signing platform
Some tool links are partner links. Hacking Demand may earn a commission at no extra cost to you.
This blueprint is a version of The Revenue Analyst (data agent) ↗
Want The Revenue Analyst built for you?
Our experts build it inside your stack. Then we hand it off, maintain it, or operate it with you. Agents and programs get better with tuning, so most teams keep us on to run them. Bring this blueprint to a scoping call.
Scope this agent ↗ Get your own blueprint ↗Good questions.
Straight answers.
Will this tell us whether marketing or sales is right?
Usually neither is wrong; they are answering different questions with different filters. The agent shows both numbers next to the agreed definition and a record-level bridge explaining each difference, so the conversation shifts from whose number is right to which records need fixing.
Most of our new work comes from referrals and existing client relationships. How does the agent credit that fairly?
Referrals get their own source values in Salesforce (past client, partner network, alliance firm) and an agreed precedence rule in the dictionary, so a referral is not overwritten by a later webinar or ad touch. The digest reports referral revenue and win rate separately, and flags deals where the referral field is blank so the partner who owns the relationship can fill it in.
Can it show whether we can actually deliver the work in our pipeline?
Yes, once time tracking is connected. The agent compares weighted pipeline by practice against partner and team utilization for the coming weeks, and highlights where proposals are stalling because the partners who need to scope them are already fully booked. Staffing decisions stay with leadership; the agent supplies the data.
Every good build
starts with a conversation.
New builds, remodels, and additions, powered by AI. Let’s plan the work.