What does a Revenue Intelligence platform do?

A Revenue Intelligence platform connects the systems where your revenue is created and recorded, such as CRM, contracts, billing and accounting, and checks continuously whether they agree. It finds deals without an invoice, prices that deviate from what was agreed, forgotten indexations and customers who are starting to buy less. Every finding gets a euro amount, evidence from the source systems and an owner, so that something is done about it instead of it sitting in a report.

That is the short version. Below is what such a platform has to do in practice, what to look for when assessing one and where the limits are. For the wider context of the discipline, see the complete guide to Revenue Intelligence.

The six jobs of a Revenue Intelligence platform

1. Pulling data from your existing systems

A platform starts with connections. It pulls deals, products and customer details from the CRM (Salesforce, HubSpot, Pipedrive), invoice lines and payments from the accounting system (Exact, AFAS, Xero, NetSuite, SAP) and, where relevant, orders, hours, subscriptions or usage from other systems.

The crucial word is continuously. A one-off export shows how things stand today. A connection that refreshes every night or every hour shows what is going wrong tomorrow. Which systems to connect for this is covered in which systems to connect for Revenue Intelligence.

A good platform asks for as few permissions as possible. Read access is enough to find leaks. Write permissions are only needed if the platform also carries out actions, and then they should be scoped and approved per action.

2. Connecting records to each other

This is the work that makes the difference between a dashboard and a check. The platform has to know that deal 4471 in the CRM belongs to contract K-2023-118, to customer account 20455 in the accounting system and to the seven invoice lines posted against it since March.

In an ideal world every system shares a unique customer number and invoices carry the order number. In practice those keys are often missing, filled in incorrectly or different in each system. A platform therefore has to be able to connect in several ways: on fixed keys where they exist, and on combinations of name, company registration number, amount and date where they do not. It also has to show how certain a connection is. A match on order number is something different from a match on a similar-looking company name.

3. Comparing with what should have happened

Once the records are connected, the platform can set the expectation beside reality. It does this with two kinds of checks.

Rules. Checks you can write down:

• A won deal has an invoice within a set period.
• The price on the invoice equals the price in the deal or contract, within the agreed discount.
• A temporary discount stops on its end date.
• After the indexation date the rate rises in line with the clause.
• The number of units invoiced equals the number of units delivered.

Patterns. Deviations that no fixed rule captures:

• A customer who ordered monthly for years and has now ordered nothing for three months.
• A customer who pays more or less than comparable customers for the same product.
• A salesperson whose deals consistently close at a higher discount than colleagues'.

Rules find the known leaks. Patterns find the leaks you had not thought of yet.

4. Expressing every finding in euros

A list of two hundred deviations helps nobody. A list of two hundred deviations sorted by what they cost per year does. A platform calculates the difference for each finding: one-off (a forgotten invoice) or recurring (a rate set too low). It also shows how certain that calculation is.

Worked example: suppose the platform finds three things.

The one-off deal is the largest amount, but the indexation carries on every year and compounds if nobody steps in. A platform that simply puts the largest number at the top misses that distinction. A good platform shows both.

5. Turning it into a task

A finding has to land with someone. The platform assigns it: the forgotten invoice to the finance team, the indexation to the account manager, the discount that carried on to sales operations. With a status, so you can see what is open and what is resolved. Without that step, every analysis ends up as a PDF in someone's inbox.

6. Preparing or carrying out actions

The last step is the fix itself. Some fixes lend themselves to automation: preparing a draft invoice for the missing additional work, creating a task in the CRM, sending a reminder to the customer owner. Others need a human decision: do you charge a customer two years of missed indexation, or only from now on?

The principle should be that the platform proposes and a person approves. Certainly for anything that goes to a customer.

What a platform should not do

A Revenue Intelligence platform is not a new source of truth. It does not replace the CRM and it is not an accounting system. The accounting system remains authoritative for what was invoiced, the CRM for what was sold. The platform is the layer that compares them.

Nor should it change invoices, create credit notes or contact customers on its own initiative. A platform that can do that without approval has more permissions than the job requires.

And it is no substitute for clear agreements. If nobody knows which discounts are allowed, the platform can only report that discounts differ, not whether that is justified.

Platform, tool or operating system?

Not every solution does all six jobs. Many tools do one or two: a forecasting tool reads the CRM and predicts, a billing tool invoices subscriptions and reports churn. A platform covers the whole chain, from reading to acting.

Some vendors go a step further and speak of a Revenue Intelligence OS: a layer that not only checks, but also drives the automations and agents that carry out the fixes. What the difference is and when it matters is explained in what a Revenue Intelligence OS is.

What to look for when assessing a platform

Use this list when you talk to a vendor, or when you decide whether to build it yourself with Power BI and spreadsheets.

1. Which systems can it read? And by which route: a standard connector, a custom API connection or manual uploads?
2. How does it connect records? Ask about the approach when customer numbers are missing. Ask whether it shows how certain a connection is.
3. Which checks are built in? Won without an invoice, price deviation, indexation, quantities, expired discounts, declining customers. How many of these do you have to build yourself?
4. Can it read contracts? If your most important terms live in PDFs, this is decisive.
5. How is a finding priced? One-off versus recurring, and on what basis.
6. Where does a finding end up? In a report, or as a task with a person?
7. Which permissions does it ask for? Read access for detection, write permissions only per approved action.
8. Where is the data stored? Within the EU, with a data processing agreement.
9. How is it priced? Per user or per company. Revenue checking needs sales, finance and the board all to be able to look in, so per-user pricing works against it.

How AutoMaat's platform handles these jobs can be seen screen by screen on the system page.

How it works in practice

A typical sequence after connecting looks like this.

The first run. The platform reads the history and finds a backlog of old differences. This is often the moment with the most findings, because nobody has ever looked systematically before. Some of them turn out to be justified (a deal deliberately not invoiced, a discount extended verbally). You record those exceptions so they do not keep coming back.

The clean-up phase. The largest amounts are picked up first. Forgotten invoices are sent after all where the contract allows it. Rates are corrected before the next invoicing run.

The continuous phase. After that, the value shifts from recovering to preventing. A new deal without an invoice is reported within days, not after a year. An approaching indexation date comes in as a task before the January invoice goes out.

That last phase is where a platform differs from a one-off check. An audit finds what has gone wrong. A platform catches what is about to go wrong. How the underlying technology works at each step is covered in how Revenue Intelligence works.

Frequently asked questions

What is the difference between a Revenue Intelligence platform and a BI tool?

A BI tool such as Power BI shows what you build yourself. You have to come up with the questions, the joins and the checks. A Revenue Intelligence platform comes with the revenue leakage checks built in and ends in a task with an owner, not in a chart.

Does a Revenue Intelligence platform need write access to my CRM or accounting system?

Not to find leaks. Read access is enough. Write permissions are only needed for actions such as creating a task or a draft invoice, and should be approved per action.

How quickly does a platform find something?

The first run over the history usually produces findings straight away, because a systematic comparison has often never been done. How many and how large depends on your company. There is no reliable average for that.

Can I do this in a spreadsheet?

For a one-off comparison of deals and invoices, yes. Continuously it becomes difficult: exports go out of date, the links between records have to be rebuilt every time and nobody has the time to repeat it every month.

Which companies is a platform worthwhile for?

Mainly B2B companies with hundreds of customers or more, recurring revenue or customer-specific pricing agreements, and several systems between deal and payment. Below that threshold, a periodic manual check is often enough.