Dynamics 365 and Microsoft Fabric: from customer data to better customer follow-up

The customer looks happy in the CRM. The purchase history tells you more.

Picture a customer who has been with the business for several years. The last meeting went well, and there is an open opportunity in Dynamics 365. The account manager plans to follow up with a new quote.

At the same time, the finance system shows that the customer is ordering less often. The delivery system shows more delays, and customer service has logged repeated enquiries about the same problem.

This is an illustrative example. Each department holds information that makes sense on its own. Taken together, it can change what the account manager should do first.

In a solution where this data is connected in Fabric, the business can identify customers with both falling purchases and unresolved delivery problems. The result can be made available in the customer view in Dynamics 365, together with the information that explains why the customer should be followed up.

The account manager can then look into the deliveries and have a clarifying conversation before the quote goes out. Falling purchases do not necessarily mean the customer is about to leave. They may reflect seasonal variation or changing needs. The combined information makes it easier to ask the right questions.

What Dynamics 365 and Fabric bring to each other

In Dynamics 365 Sales and Customer Service, people work with customers, contacts, opportunities, activities and cases. This is also where the context of the customer relationship lives: who follows up the customer, what has been agreed and what happened in the last conversation.

Fabric can combine this information with data from finance, deliveries, product usage and other sources. The business can prepare history, calculate metrics and analyse patterns across systems. Microsoft's Dataverse link makes CRM data available for this work in Fabric. See Microsoft on the link between Dataverse and Fabric.

CRM data therefore makes the analysis easier to interpret. A drop in orders is easier to understand when it is seen alongside a contract that has ended or a planned change at the customer. Fabric, in turn, can give the CRM user a broader basis for decisions than the customer dialogue alone.

The follow-up still happens in Dynamics 365. That is where people can record what the customer actually needed, which actions were agreed and whether the issue was resolved. This information can later feed into the analysis of which signals and actions proved useful.

Assess the profitability behind the opportunity

A large opportunity can look attractive when judged on expected revenue. Its profitability may depend on factors the salesperson cannot easily see: discounts, returns, freight costs or extensive service work.

In a second illustrative scenario, Fabric is used to combine historical revenue and relevant costs with customer and product information from the CRM. Finance and sales first have to agree which costs to include and how to allocate them.

The salesperson can then see an agreed measure of the customer's historical profitability in Dynamics 365. That can support adjusting price, delivery terms or scope before a new quote is sent.

Historical margins do not determine what the next deal will deliver. They do, however, make the assumptions behind the quote easier to assess.

Make additional sales more relevant to the customer

A third example is finding customers who could benefit from an additional service. Purchase history can show which products the customer already has. Usage data can point to a need for more capacity or support. The CRM can show that the customer recently discussed an expansion with the salesperson.

When the information is combined, the business can produce a reasoned list of customers worth considering for follow-up. The criteria can be simple business rules. More advanced models can be considered if the data and the need justify it.

In Dynamics 365, the salesperson should see why the customer has been suggested, which information the assessment is based on and when it was last updated. The salesperson can then use their knowledge of the customer to decide whether getting in touch makes sense.

Sometimes it is right to wait. A customer with an unresolved complaint may need help before another sales conversation.

How the insight reaches day-to-day customer work

To benefit from the combination, the business has to decide how the insight will be used. A calculated value in Fabric is only a starting point.

One option is to show selected Fabric data in the CRM through virtual tables in Dataverse. They make external information available in the app without copying the whole data set into ordinary CRM tables. Microsoft describes, among other things, how this can give account managers a combined overview in a model-driven app. See Microsoft's example architecture for a customer overview.

Another part of the solution can be a workflow that reviews new signals at regular intervals and creates a follow-up task in Dynamics 365. Power Automate can be part of such processes. You need to define who gets the task, what it should contain and how to avoid repeated alerts about the same issue. There is more on automation in the sales process in Automate the sales process with Power Automate and Dynamics 365.

Both the display and the actions have to be set up as part of the solution. The Dataverse link to Fabric does not create tasks or change customer data automatically. The virtual tables from Fabric are read-only. Tasks and changes in the CRM can be handled by the user or by a workflow set up with the necessary permissions. See Microsoft on using Fabric insights in apps and workflows.

Settle data links, access and responsibility

For the examples to work, the information has to be linked to the right customer. The customer number does not have to be the same in every system, but the link between them must be reliable. You also need to decide whether the analysis covers a single department, one company or the customer's whole group. Why duplicates in the CRM make this harder is covered in CRM data quality.

The definitions must be clear. What counts as falling purchases? Which costs are included in profitability? How old can the information be before it loses its value for follow-up?

Access has to be considered for the whole solution. The fact that information exists in Fabric does not mean everyone using the CRM should see it. Virtual tables from Fabric do not support Dataverse row-level or field-level security. Security roles can control who sees the table, but not which rows or fields each person sees. Access control therefore has to be designed specifically for the information that will be shown. See Microsoft's security considerations for this architecture.

Finally, decide who will act on the signals. Each prioritised signal should have a recipient who understands why it matters and what they are expected to do.

Start with one kind of follow-up you want to improve

A good place to start is a concrete situation your people already know: customers with falling purchases, quotes with uncertain profitability or opportunities for a relevant additional service.

Select the information you need and try the solution with a small group of users. Check whether the signals are understandable, whether they lead to useful follow-up and whether people get information they would otherwise spend time looking for. Record false positives as well, so the criteria can be improved.

Fabric becomes particularly relevant when such analyses need data preparation and history from several systems, and the results are to be reused. Some needs can be met with existing reporting or integration. In Microsoft Fabric and Dynamics 365: when do you need a shared data platform? we look more closely at the platform choice, the costs and the technical prerequisites.

The value of the combination shows up in customer work: an account manager spots a problem earlier, assesses a quote better or gets in touch with a proposal that fits the customer's needs.