Agents entered the seller’s working day

At Ignite on 21 November, Microsoft described new AI capabilities in Dynamics 365 Sales. Sales Qualification Agent would research leads and help sellers prioritise. Sales Order Agent would support the order process, while Copilot could prepare meetings, summarise communication and propose follow-up. AI was moving from a separate writing tool into specific CRM steps.

The connection to service and marketing mattered. A lead could originate in a customer journey, and meeting preparation could use earlier contact and open cases. When a promise required technical follow-up, field service needed the correct agreement context. Automation without a shared customer identity and clear statuses would instead generate more proposals from an incomplete history.

Meeting preparation was a particularly suitable review point. A seller could compare the agent’s summary with the latest email, open cases and agreed activities before the meeting. A missing source could be corrected without affecting the customer. The test also revealed whether customer data was connected, current and understandable across teams.

Action required a defined mandate

Microsoft’s wider agent launch also covered Case Management Agent and Scheduling Operations Agent, suggesting a chain from qualification to case and visit. The announcement, however, combined previews and planned rollouts. Organisations needed to verify actual availability in their own region and tenant before including a capability in a benefits estimate.

For each agent, the process owner had to distinguish retrieving information, proposing a next step and making a change. An inaccurate meeting summary could be corrected before use. A poor priority or order could immediately affect a customer and capacity. Roles, logs and boundaries on automatic action therefore belonged in process design rather than being added after a pilot.

Customer productivity introduced a different measure

Gartner published research on 1 November cautioning against a narrow focus on internal productivity. Organisations should also use AI to make customers’ lives easier. In sales and service, that could mean fewer repeated questions, more accurate information before a meeting and a clearer handover from agreement to delivery.

A November experiment could follow one lead category through the whole flow. The team could measure time to the first relevant response, meetings with missing history, manual corrections and customer waiting during handovers. The agent would then be judged against a real customer process. A higher volume of activity alone was not evidence of better sales or service.

The customer measure could be supplemented by a qualitative review of a small sample of handovers. A seller, service representative and dispatcher could read the history together and identify what they had to ask again. Improvement would then address missing data and unclear responsibility as well as the wording produced by the agent.

Sources

Microsoft Dynamics 365 Blog: New AI capabilities in Dynamics 365 Sales, 21 November 2024.

Microsoft Dynamics 365 Blog: Transform work with autonomous agents, 21 October 2024.

Microsoft Dynamics 365 Blog: 2024 release wave 2 launches, 29 October 2024.

Gartner: The Customer Side of AI-Driven Productivity, 1 November 2024.

For discussion

Which step in customer follow-up can be automated so both the customer and employee spend less time?