Decagon's Missions Target Long-Running AI Customer Workflows

By Nathan Brooks

Decagon introduced Missions on November 12, 2026, a product designed to let AI agents pursue a defined customer objective across days, weeks or months. The workflow spans multiple conversations, systems and participants, according to the company, rather than ending with one interaction.

Decagon says Missions should replace manual status tracking and handoff coordination with ongoing refinement of the customer journey. It cites three multi-step journeys that can generate meaningful revenue:

• Prior authorization for an MRI
• Small-loan approval
• Lead qualification

A team starts each Mission with three inputs:

• A goal stated in plain language
• The personas the agent should engage
• Additional guidance for the agent

According to Decagon, the agent chooses subsequent actions around that objective. It can connect threads involving different people, contact each participant through an appropriate channel and wait when a third party must complete a step before the process can continue.

What changes for customer-experience teams?

Decagon’s prescription-refill example shows how a Mission can carry one service outcome through several operational layers:

• Notify the patient
• Contact the provider when authorization is required
• Run claims again on a payer portal
• Confirm pickup details with the patient

The company says Browser Actions also let agents complete back-office work, including forms. Decagon places these activities in “one goal-directed workflow,” rather than treating each interaction as a separate job.

For customer-experience teams, the change is practical: Decagon expects them to devote less effort to tracking status and coordinating handoffs. The journey, rather than the individual chat, becomes the unit that agents manage and teams refine.

Each Mission case records completed steps and new information that could alter the plan. Decagon says the case guides the next action, proactive outreach and retained context for later work. Before acting, the agent rechecks both the case and customer Memory to determine whether the existing plan still fits.

Memory supplies preferences and relevant context from other conversations. Decagon’s examples include honoring a request not to receive calls after 3 p.m. and recognizing that the same customer may be onboarding a new product while troubleshooting an unrelated issue.

The recheck also handles a customer who changes the expected sequence. If an agent qualifies a lead and arranges a follow-up, but the customer independently schedules that follow-up by email, the agent adjusts its course. Decagon draws a deliberate distinction: the Mission case tracks progress toward the business result, while Memory keeps individual preferences and concurrent interactions visible.

That separation matters because long-running automation can obscure where a delay originated. Decagon identifies attribution as a central difficulty when work stretches across many interactions. If a prescription refill is delayed, teams may need to determine whether the problem began with initial outreach, the provider handoff or insurance verification.

Decagon builds every Mission around an “Agent Operating Procedure,” or AOP. Teams can write, edit and refine that logic, inspect the agent’s decisions and test individual parts of the configuration to identify what needs improvement. The company says this testing addresses attribution across the journey rather than at one interaction.

Decagon also says Duet and Duet Autopilot provide recommendations aimed at the full customer journey. Those recommendations extend beyond isolated interactions, matching Missions’ longer operating horizon.

Decagon’s launch links execution and diagnosis in one product model. Agents carry an objective across participants and systems, while cases, Memory and AOP-based testing give teams tools to review and revise the path taken. For enterprises considering longer-running automation, the practical test is whether those controls can reliably expose weak handoffs and improve outcomes as the number of interactions grows.