1. Fix Intake Before Enabling AI Features
Suggestions, Recommendation Cards, and the SR Creation Assistant all depend on structured, consistent data at intake. If your required fields aren't enforced and your SR data is messy, AI won't perform reliably downstream. This is a data problem, solve it first.
2. Connect Your Knowledge Sources Before Activating AI Agents
The SharePoint and Web Crawler connectors are available now, and sequencing matters. If you activate AI agents before connecting your existing knowledge repositories, those agents are working with an incomplete picture. Get your sources connected first, then layer AI agents on top of a complete knowledge base.
3. SR Summarization is Now Configurable - Time to Review Your Current Setup
Most teams that have summarization enabled have never tuned it. Now that it's a configurable AI Agent, you have real control over how summaries are generated and presented. Don't leave it on default. Align it to how your team actually uses SR information and what they need to see.
4. Define Your Resolution Paths Before Enabling AI Action Suggestions
This is a larger-scale feature that requires setup to deliver value. Teams that enable it without first mapping their most common SR types and resolution workflows will get generic suggestions that agents learn to ignore. Do the mapping first, even for your top five SR categories, and the feature becomes genuinely useful.
5. Treat Help Desk AI Separately From Service Center AI
Help Desk now has its own AI summarization capability. It's tempting to configure it the same way as Service Center, but the audiences, data structures, and use cases are different. Internal HR requests look nothing like customer-facing service requests. Configure them independently and you'll get meaningfully better results from both.