💬 AI Agents
AI agents and workflows are now easier to configure, more adaptable to your users, and smarter to test — giving you extra control and confidence at every stage of automation.
- User segments allow you to tailor your AI agent’s replies based on user type. Customize tone, escalation logic, or article suggestions depending on who the AI is speaking with, so each user gets the most relevant experience. User segments unlock endless possibilities —set your VIPs to get escalated to humans immediately, ensure admins get answers tailored to their access level, or send self-serve links to first-time users. Learn more.

- We’ve released a powerful new type of workflow step that lets you direct your AI agents using dynamic, natural-language instructions. Instead of building out complex, branching flows, you can now pull in live data from tools like Google Sheets, outline constraints, and let the AI reason through the best path — all in one flexible step. This dramatically reduces the time needed to build and maintain workflows, replacing dozens of static nodes with a single, intelligent guide.

- Workflow previews are now enhanced with more transparency and configurability. Users can now explicitly select which workflows they want to run as part of their preview, as well as what version of the current workflow they want to run the preview on.

- In workflow testing, users now have more control over which tickets to evaluate in bulk — including options like random sampling of true positives and negatives, and the ability to set test parameters such as interaction types and target workflows.

- We’ve enhanced our workflow selection to include a more detailed rationale, as well as a “Get AI suggestions” feature. This improves reliability and makes it easier to build effective automations — no expert prompting required.

- AI email agents now includes even more configuration settings — from triggering a workflow based on your own custom email event to assigning the ticket explicitly to an AI agent.

- Use the new Chat Responses API to integrate your own frontend support chat interface with the AI Chat Agent backend, and use the new Edit Endpoint option to test custom API endpoints.
🛠️ Copilot
You can now test Copilot email automations with a human in the loop before taking action. When reviewing AI-generated responses, you'll have visibility into which workflow steps were triggered — making it easy to validate behavior and catch issues early.

📆 Workforce management
We’re expanding what scheduling rules can do — giving you more flexibility, better constraint handling, and clearer feedback during setup. These updates help lay the foundation for our larger roadmap initiative for optimal schedule generation. Scheduling rules are currently available by request only on Pro and Enterprise WFM plans. Reach out to your Assembled CSM or support@assembledhq.com to learn more.
- Add minimum shift constraints, like “agents must work at least 35 hours/week,” to ensure coverage goals are met — complementing the existing max-hour rules.

- Build more powerful schedules with recurring events. Now you can specify that an event should occur weekly or monthly, and Assembled will find the optimal time to place the event when generating schedules.

- We now support multi-agent, multi-time-range assignments in a single rule, so you can avoid creating duplicate logic for rotating teams or overlapping groups.
- Now you can schedule events conditionally based on the duration of another event — add breaks after long chat shifts or wrap-up time after extended work blocks.
- When creating rules, you’ll also see inline validation that flags any issues, helping you troubleshoot in real time.
We're making workforce management even easier to set up, measure, and automate with more updates for scheduling and forecasting.
- Quickly configure schedules by uploading working hours via CSV. This is great for large teams, rapid onboarding, or changing agent hours across different time periods in bulk.
- Get clearer insights with two new metrics on Forecasted vs. Actuals, within the Support Volume report:
- Actual required staffing, which compares your forecasted vs. real staffing needs based on actual contact volume
- Actual occupancy, which shows how much time agents spent handling contacts vs. available — helping identify under- or over-utilization
- New API endpoints for time off auto-approvals, forecast adjustments, and forecast outliers are now available!