[object Object]

A customer service director inherits a Dynamics environment with 40 case statuses, six SLA configurations, and three different routing patterns built by three different consultants. The system technically works; the team avoids it because the cognitive load of using it is higher than a shared inbox. The fix is not more configuration; it is less. Customer Service rewards simplicity at the data model level and discipline at the routing level.

Case Management Basics

Cases have a lifecycle: New to In Progress to On Hold to Resolved. Customize with stages, SLAs, escalation rules. Do not over-customize initial status — simpler is easier to report on. The minimum viable status set is five values; anything beyond that should be justified by a specific report.

Recommended case status set:
- New (just created, not assigned)
- In Progress (agent owns it)
- Waiting on Customer (paused for customer input)
- On Hold (waiting on internal team)
- Resolved (closed positive)
- Cancelled (closed without resolution)

Six values. Resist adding more.

Omnichannel

Omnichannel for Customer Service adds voice, chat, SMS, WhatsApp, messaging apps on top of email and phone. Unified agent desktop shows all conversations in one place. The unified desktop is the value proposition; agents who can switch channels without switching tabs handle 30 to 40 percent more conversations per shift.

Channel coverage decision:
- Voice: required for traditional support
- Chat: required for digital-first customers
- Email: still 30 to 50 percent of volume in most B2B
- SMS: useful for outage notifications and confirmations
- WhatsApp/Messenger: regional, depends on customer base

Enable channels based on customer behavior, not vendor demos. A WhatsApp channel that gets 5 messages a week is not worth the agent training.

Unified Routing

Skill-based and assignment-rule-based routing. Agents get work matched to their expertise, availability, and workload. Configure rules; measure impact; tune. The routing engine evaluates rules in order; the first match wins, so rule order is the most important configuration choice.

Standard rule order:
1. VIP customer routing
2. Skill-based routing for technical issues
3. Language-based routing
4. Geographic routing for regional teams
5. Catch-all queue

Build a catch-all at the bottom or cases that match no rule sit unassigned forever.

Knowledge

Knowledge articles surface inline in cases. Agents see relevant articles ranked by relevance. Customer Service Copilot uses knowledge plus case data to draft responses. The relevance ranking depends on article tagging; under-tagged knowledge degrades search results.

Article metadata to maintain:
- Title (clear, searchable)
- Keywords (synonyms, common phrasings)
- Category (matches case routing)
- Product (if multi-product)
- Last reviewed date (drives age-based reranking)

Run a quarterly knowledge review; articles older than 12 months get rated for currency or retired.

Customer Service Insights

Dashboards on case volume, resolution time, CSAT, backlog. Publish KPIs weekly. An insight no one reads is not an insight. The leading indicator metric is backlog growth; if cases create faster than they resolve over a sustained period, the team is sliding behind regardless of CSAT.

<fetch aggregate="true">
  <entity name="incident">
    <attribute name="incidentid" aggregate="count" alias="cnt" />
    <attribute name="statecode" groupby="true" alias="state" />
    <filter>
      <condition attribute="createdon" operator="last-x-days" value="7" />
    </filter>
  </entity>
</fetch>

Compare the open count across weeks to spot the trend.

Productivity Tools

Smart Assist, Macros, Customer Summary, and the Productivity Pane reduce agent toil. Configure macros for the top three repetitive tasks (account verification, password reset, status update). A macro saves 30 seconds per use; at 100 uses per day across the team, that is meaningful time back.

Escalation Patterns

Escalation should change priority, change assignment, and notify a supervisor. Build a single escalation flow that handles all three; do not build per-team escalation logic. The pattern that works is a flow triggered when SLA warning fires, which checks priority, escalates if appropriate, and notifies the on-call supervisor.

What to do this week

Audit your case status list and consolidate to six values. Build a single escalation flow and retire any per-team duplicates. Run the backlog trend report and share weekly with the service leadership team.

[object Object]
Share