CRM · Salesforce data quality for AIYour Salesforce looks clean. Ask your ledger.
Salesforce is where many AI ambitions start: scoring, forecasting, an agent on the pipeline. It is also the system most likely to disagree with finance about who your customers are and what they paid. We compare your Salesforce exports with the systems beside it, matching customers, deals and invoices across them, before anyone builds on it.
Starts with a 30-minute call. Commits you to nothing. Then exports only: no login, no connected app, nothing installed.
Source of truth
What Salesforce gets right.
Every system is built to be right about something. Knowing what Salesforce owns, and what it doesn’t, is where a cross-system check starts, and where the Truth Score™ (a score out of 100 for how far your data can be trusted) begins to take shape.
- 01
The pipeline
Open opportunities, stages, close dates and forecast categories. Nobody else holds this.
- 02
Who owns each account
Account and opportunity owners, territories, and the sales hierarchy.
- 03
The commercial relationship
Contacts, roles, activity history and how the deal was won.
Revenue. A Closed Won opportunity is a sales record, not an invoice. Billed and collected revenue lives in finance, and the two drift apart through discounts, credits, renewals and deals that were never billed.
Inside the system
Common Salesforce data problems for AI.
What goes wrong inside Salesforce itself, with the objects and fields where it shows. Each one is checkable from an export.
- 01
Duplicate Accounts and Contacts
Lead conversion, list imports and manual entry create the same company two or three times. A model reasoning over relationships, or an agent looking up a customer, finds the wrong one.
Account.NameAccount.WebsiteContact.EmailLead.ConvertedAccountId - 02
Open opportunities that are already dead
Deals past their close date still sit in open stages. Forecasting and win-rate models learn from a pipeline that stopped being real months ago.
Opportunity.StageNameOpportunity.CloseDateOpportunity.LastActivityDateIsClosed - 03
Stages and picklists that changed meaning
Sales processes get redesigned; old stage values and record types stay in the history. Records from 2023 and 2026 no longer compare.
StageNameForecastCategoryNameRecordTypeIdOpportunityHistory - 04
Owners who have left
Accounts and open deals stay assigned to inactive Users. Ownership is one of the first signals routing and scoring models use.
Account.OwnerIdOpportunity.OwnerIdUser.IsActive - 05
No key that finance shares
Many orgs never store the ERP or billing customer ID on the Account, so matching to invoices falls back to names, which are spelt differently in each system.
Account.ERP_ID__c (custom)Account.NameAccount.ParentId
Between systems
Salesforce looks clean on its own. The gaps are next door.
The problems that sink AI sit between Salesforce and the systems beside it: the same customer, person or transaction recorded differently in each. These are the gaps to test for each pairing.
CRM
Salesforce
Source of truth forThe pipelineWho owns each accountThe commercial relationship
Salesforce ↔ Finance and billing
NetSuiteXeroSage IntacctQuickBooks
- Closed Won deals with no matching invoice
- Accounts with no matching customer in the ledger
- Opportunity Amount that never matches what was billed
Salesforce ↔ Marketing automation
- The same person as a Lead and a Contact
- Campaign influence the marketing system can’t reproduce
Salesforce ↔ HR and payroll
- Accounts owned by people who have left
- Territories that don’t match the org chart
Salesforce ↔ Support
- Tickets that tie to no Account
- Support entitlements that don’t match the Contract
What to export
What we need from you, in plain English.
Salesforce’s own Data Export (in Setup) or ordinary report exports both work. We send the exact field list after the call. No login, no connected app, nothing installed.
- Exports only, no access to your systems
- NDA and data processing agreement first
- The full list of sub-processors comes with our data processing agreement, before any data moves
- Raw exports deleted after the engagement; we keep quality metrics, not your records
AI that depends on it
AI use cases that lean on Salesforce.
If one of these is in a proposal or your plan, Salesforce is one of the systems we compare. Each page sets out what the use case needs and where it breaks.
- Revenue Revenue forecasting The pipeline is the input. Stale stages and slipping close dates set the ceiling on any forecast.
- Revenue Lead scoring Win and loss history on Opportunities is the training target. Duplicates and missing outcomes corrupt it.
- Revenue Customer churn prediction Account and Contract records have to join to billing and support to see a customer whole.
- Revenue Pricing and rebate optimisation Quoted prices and discounts on Opportunity Products only mean something once they match what was invoiced.
- Operations Customer service automation An agent answering customers needs Cases tied to the right Account and Contract.
How a Truth Score Read tests Salesforce
From your exports to a signed verdict.
Within 5 working days of complete exports. A fixed fee, agreed before we start, that never moves with the verdict.
- 01
Export Salesforce, and one system beside it
The Salesforce exports listed above, plus at least one neighbour, such as finance and billing. Under NDA, using the field list we send.
- 02
Match across systems
Our engine matches each opportunity and its related records across the exports by ID, name, email and domain, including the ones spelt differently in each system.
- 03
Check by hand, then score
Every red flag is checked by hand against sample records before sign-off. Then the Truth Score: each pillar shows what was measured and what wasn’t, and pillars with too little signal are marked, never guessed.
- 04
Get the verdict in writing
A signed memo: the red flags between Salesforce and the other systems, a Leverability read on the AI you’re considering (which AI ideas you can use now, after fixes, or not at all), a verdict (sign, fix first or walk away) and a short fix list your own team can work from.
What you get back
See what comes back: a real sample memo.
One fictional B2B software company, Example Co: 72,837 records in 7 exports from its CRM, finance and HR systems, run through the same engine and the same checks as a paid Truth Score Read. Each system passes its own checks. Matched across systems, only 13% of deals tie to an invoice.
Questions
Salesforce and AI, answered plainly.
Anything else, ask on the 30-minute call. It commits you to nothing.
Book the callCan you assess Salesforce without logging in to our org?
Yes. We work from exports only: one person on your side runs standard report exports or Salesforce’s Data Export, under NDA, using the field list we send. We never need a login, a connected app or anything installed.
Why compare Salesforce with finance at all? Isn’t the CRM enough?
On its own, Salesforce usually looks tidy. The problems that sink AI sit between systems: deals marked won that were never invoiced, accounts finance has never heard of, owners HR says have left. You can only see those by matching records across both.
We use Einstein or Agentforce. Does that change anything?
No. Any AI that reads your org inherits what is in it. The question is the same whoever supplies the model: does the data it needs exist, and does it agree with the rest of the business?
How far back should the opportunity export go?
Three years is a sensible default for forecasting and scoring, so stage changes and seasonality show. If your sales process was redesigned, include records from both sides of the change; that is where the inconsistency shows.
Do you fix what you find in Salesforce?
No. We never fix what we find and take no referral fees from anyone who does. You get the findings and a fix list your own admin, or any partner you choose, can work from.
Find out what your Salesforce data can carry. Before the AI does.
Your Salesforce export and one other system in. The gaps between them in writing, within 5 working days of complete exports. Starts with a 30-minute call. Commits you to nothing.
Starts with a 30-minute call. Commits you to nothing.