Agentforce readiness · Sales, Service and Marketing Cloud
Four measurements against reports you already have tell you whether an agent deployed on this data would forecast against dead pipeline, treat one buyer as three accounts, and read half your customer contact as silence.
Runs in this tab. No org access, no OAuth, no upload.
Stale pipeline
41.3%
1,208 of 2,924 · healthy < 25%
Duplicate identity
18.7%
602 of 3,219 · healthy < 10%
Activity capture
0.48
per open opportunity · healthy > 2.0
Source attribution
22.1%
top value "Web" · healthy < 25%
Fix first, then deploy. Three of four measures sit outside range. At an average deal of €45,000 the stale set alone is roughly €54M of pipeline an agent would read as live.
Drop your own exports here · opportunities, accounts, leads, activities as CSV. Any subset works; each file replaces the matching measure.
Your deliverable
This is the document the €500 engagement produces, filled in with your own numbers. Page two carries the query behind every figure so your admin can reproduce it without me.
Standard reports exported by your admin. No login, no permission set, no change to the org.
-
Build the case
Untick anything you would not fund this quarter. The figures below are what you take to whoever signs the budget.
Elapsed
-
Investment
-
Pipeline made trustworthy
-
Verdict after
-
Leave the address empty to mail it to yourself from your own client.
Page two lists the SOQL behind every figure. One email, no sequence.
Or send it to me
Your four measures travel with the message, so the reply starts from what is actually broken rather than from a discovery call. If the answer is that you do not need the work, that is what the reply will say.
What is measured
Your admin exports four standard reports. Nothing is installed, no permission set is created, no field is touched. The parsing happens in this browser tab and the file never leaves the machine.
Open opportunities with a close date in the past, or no logged activity in thirty days. An agent asked to forecast cannot tell these from live deals, so it forecasts against them.
Accounts sharing a website domain with at least one other account. Account-level reasoning breaks here first: one buyer read as three companies, three separate next-best-actions.
Logged emails and meetings per open opportunity. Whatever is not logged does not exist to the agent, and silence is the one signal it reads as disengagement.
The largest single Lead Source value as a share of inbound. When one value swallows the field, the agent has no basis to reason about where revenue comes from.
Proof
The deliverable is two pages. Page one is the verdict. Page two is how each figure was derived, written so your admin can reproduce it without me. If you later hire someone else, the document still works.
Run this in the developer console of your own org and compare the count against the number above. If they disagree, the finding is wrong and I say so.
This check does not review your automation, your security model, your Apex or your integration design. It answers one question: can the data underneath carry an agent. Those other questions are real, and they are separate engagements with separate prices.
It also does not predict adoption. Gartner projects forty percent of agentic projects cancelled before production through 2027, and most of that is organisational rather than technical.
Engagements
No retainer, no hourly billing, no discovery phase billed as a project. The same senior consultant on the first call does the work.
Data readiness check
€500
3 days · no org access
The four measures above against your full data set, a verdict of A, B or C, and a costed order of work. Credited in full against an engagement booked within sixty days.
Foundation build
€3,000 per fault
2 to 4 weeks per fault
Fix what the check found. Pipeline hygiene, account deduplication with a survivorship rule, activity capture enabled, attribution rebuilt. The two this check finds most often are sync at €3,000 and attribution at €5,500, which is €8,500 together. Documentation and sixty days of support included.
Full architecture
€25,000 and up
Custom timeline
Agent design across Sales, Service and Marketing Cloud, with the data model, the guardrails and the handover that lets your team own it by day ninety.
Fit
Thirty minutes. Bring the symptom, leave with a straight answer about whether it is a data problem, an architecture problem, or neither.
Serhii Skrypnyk · RevOps Architect · 7 Salesforce certifications · on the platform since 2018
Questions
Four measures against standard Salesforce reports: stale pipeline as a share of open opportunities, accounts sharing a website domain, logged activities per open opportunity, and the largest single Lead Source share of inbound. Healthy is under 25 percent, under 10 percent, above 2.0 and under 25 percent respectively.
No. Your admin exports four standard reports and the parsing happens inside the browser tab. There is no login, no OAuth, no permission set and no file uploaded to any server.
One or two of the four measures sit outside the healthy range. The foundation is repairable inside a quarter, and each failing measure changes what an agent concludes about the pipeline it reasons over.
An agent asked to forecast cannot distinguish a dead opportunity from a live one. At a 41 percent stale rate and an average deal of €48,000, roughly €58 million of pipeline would be read as live.
The paid check runs against the full data set rather than an export sample, and produces a two-page document where page two carries the query behind every figure so the client's admin can reproduce it independently. It costs €500 and is credited against an engagement booked within sixty days.