Microsoft Copilot doesn’t reason from intent; it reasons from data. Specifically, from what’s stored in Dataverse. So, if that data is incomplete, in that customer interactions are missing, fragmented or never written back into the system, then Copilot’s outputs will be too.  

For most organizations, that is exactly the situation. Around 60–70% of customer conversations never make it into CRM. Calls, WhatsApp messages, Teams chats, emails – the richest source of customer intelligence in the business — simply disappear. Agents lack context, customers repeat themselves, and Copilot, sitting on top of an incomplete record, fills in the gaps as best it can. 

Dataverse is only as good as what gets written into it  

Dataverse is the structured data layer underpinning Dynamics 365 and the Power Platform. In principle, it holds the complete customer record that Copilot and other AI tools operate from. In practice, it reflects whatever has actually been captured, which, for most organizations, is a partial version of reality. 

When interactions are missing from Dataverse, the consequences are practical and immediate. 

• Customer timelines are incomplete 

• Activity histories have to be reconstructed after the fact 

• When Copilot generates a summary, a prompt or a recommendation, it is working from that same incomplete record 

There’s a missing layer between where conversations happen and where data gets stored. Most organizations have three things in place: systems where work happens (Teams), systems where data is stored (Dynamics 365), and AI tools sitting on top trying to interpret that data (Copilot).  

But businesses often lack a reliable connection between them. 

Conversations happen across voice, messaging, email and internal collaboration, but only a fraction are automatically captured, structured and written back into Dataverse. Agents fill in what they can manually, but much of it never gets recorded at all. 

What changes when the data is complete? 

Solgari is the customer engagement app that captures all the conversations inside Teams and Dynamics 365, recording every channel and writing it all back into Dataverse automatically – as each conversation ends, with no integration project required. Voice calls are transcribed and scored. SMS and WhatsApp conversations are captured as structured records. Action items and commitments from every channel are extracted and stored. All of it is available to Copilot on day one. 

 When done right, businesses start to see: 

• Customer histories becoming complete, not reconstructed – everything from emails and meetings to phone call transcripts, sentiment scores and WhatsApp conversations in one timeline 

• Agents working from context, not fragments – with action items and commitments from every channel surfaced automatically 

• Compliance moving from best effort to auditable record 

• Copilot operating on the full interaction, not a summary of a summary 

• Repeat contacts falling, first-contact resolution improving, and cost to serve coming down 

With Marketplace and MACC alignment, organizations already in the Microsoft ecosystem can go live in days – without lengthy procurement cycles or complex integrations. 

The organizations getting the most from Copilot are the ones with the most complete data. If your Copilot deployment is underperforming, the answer is more likely in Dataverse than in the model. 

Solgari captures all conversations to make the most complete dataset possible – natively, inside the tools your teams already use.