The Pulse
Appian expands agent control tools, with more features planned
Appian says controls, orchestration and process intelligence for AI agents are available now, with more capabilities planned. It also unveiled Ask Appian, an interface for requesting work in plain language.

Appian has expanded its tools for managing AI agents in business processes, with some controls available now and further capabilities planned. The company disclosed the changes on November 17 at Appian Europe, its annual conference in London, and also unveiled a plain-language interface called Ask Appian.
Source prnewswire.com
Many controls, orchestration and process intelligence capabilities are available today, with additional features due to become available starting this year. Appian said planned features and release timing can change and are not a commitment.
Managing agents within processes
Appian describes its control plane as a way to see, measure and control agents across the platforms on which they are built. Its approach places agents inside deterministic business processes alongside people, systems and rules.
The platform is designed to assign each process step to an agent or another resource, with guardrails and shared context for agents. If an agent fails or is stopped, its task is routed to a person.
The controls are intended to let teams monitor agent activity and costs, choose which models agents can use, track policy compliance and set alerts. Appian also describes evaluations of agent accuracy and measurement against the business process’s key performance indicators.
Third-party agents can already work as steps in Appian processes through MCP or direct integration, according to the company’s product page. A registry that includes third-party agents, tracing for agents running outside Appian and support for the A2A protocol are listed as coming next.
Measuring the work
Appian agents are already measured against the key performance indicators of the processes in which they work. The company also plans estimates of cost per case and per step for Appian and third-party agents, comparisons with manual work before and after an agent goes live, and recommendations to improve each agent.
Appian said its insights can identify repeated agent steps that could be converted into deterministic workflows and business rules. In a 2026 study conducted by Appian and Harvard Business Review Analytics Services, 16% of respondents reported a high degree of measurable value from AI.
Raiffeisen Bank International intends to use the platform to automate payment inquiries. In testing, the bank combined AI agents with rules-based automation and moved predictable steps away from agents, while retaining agents for steps requiring judgment.
Claus Mitterlehner, the bank’s head of smart automation, said the hybrid approach reduced token consumption and processing time by 80%. He attributed much of the reduction to the team’s process-centric approach to orchestration.
New interface and access
Ask Appian is intended to let users describe an outcome in plain language rather than navigate enterprise software or search across disconnected systems. The interface pulls context from systems connected through Appian’s data fabric and dispatches agents to carry out work.
Customers can join Appian’s early adopter program to use the control plane and access capabilities ahead of general availability. The company advised customers to base purchasing decisions only on features generally available today.