Appian adds AI-assisted development & MCP integration

Appian has added AI-assisted spec-driven development and Model Context Protocol integration for agents to its software platform, expanding how AI tools connect with business processes and data.

The update focuses on two areas: AI agents that can work across external systems with more context, and development tools that use AI to generate application specifications from legacy software. The aim is to place AI within structured process models rather than treat it as a standalone layer.

Appian also outlined new work on agent learning. Users will be able to track agent performance and apply an agent's memory across different processes to improve decision-making. Future options are intended to let users set optimisation goals and recommend changes that can be applied safely.

Agent changes

Appian's adoption of Model Context Protocol, or MCP, is intended to let its agents interface securely with external systems. The same approach would also allow third-party agents to use Appian tools, including its data fabric, which provides unified read-write access to enterprise data.

The data fabric has also been updated with a unified metadata model. Appian said the change gives agents more context on how information is structured and linked across systems, with the goal of creating a common vocabulary for business data.

Appian is also launching a technology partnership with Snowflake. It described the arrangement as combining its process orchestration and data fabric with Snowflake's AI Data Cloud, including direct MCP-enabled integration between Appian data fabric and Snowflake.

Snowflake said the set-up would allow agents to interact directly with Snowflake Cortex AI, supporting decisions based on data held across enterprise systems.

"Enterprises don't need more AI experiments, they need AI that delivers real business outcomes on governed data," said Baris Gultekin, Vice President of AI, Snowflake. "By combining Appian's process orchestration and data fabric with the Snowflake AI Data Cloud, we're bringing intelligence directly into the flow of work. Together, we enable secure, enterprise-grade AI where agents can access trusted data through Cortex AI, act with context, and drive measurable impact across the business."

Development tools

Alongside the agent updates, Appian introduced what it calls AI-assisted spec-driven development. The feature uses AI to extract specifications from legacy applications and turn them into a visual plan covering user interfaces, data models and process flows.

AI developer agents, working under human supervision, can then complete tasks against those specifications. Appian said this approach can reduce rework while helping teams modernise older applications.

New developer MCP servers are also expected to let organisations use external AI development tools to build and update Appian applications. Appian said it would support a range of AI models so teams can work in their chosen environments.

Mike Beckley, Chief Technology Officer and Founder of Appian, linked those tools to the broader platform changes.

"Appian Composer, Agents and Appian MCP servers enable trusted agentic process orchestration and application modernisation," Beckley said. "Composer complements Appian's agentic orchestration and data fabric with new spec-driven development tools that are both conversational and iterative. Beneath the covers, Appian Composer is built on Appian's new open MCP - a model-driven representation of your complete application estate-requirements, apps, data entities, logic, workflows, security/governance rules, integrations, and multi-object dependencies-now exposed as context for developers and agents to safely evolve and optimise."

Customer example

Appian pointed to Global Excel Management, a healthcare risk management provider, as an example of how customers are applying AI in process change programmes. The company is using Appian to reshape claims workflows.

"As part of our digital transformation we are evolving our claims processes by transitioning from fragmented workflows to an enhanced level of operations using technological advancements enabled with AI features," said Pascal Tanguay, SVP, Global Technology Services, Global Excel Management. "With Appian, our processes will be unified. From initial intake to adjudication, our advanced technology will reduce redundant tasks and lessen complexity for our team members. This ensures that our claims processes are consistent and completed more efficiently and accurately."

Appian's broader message is that AI systems are more useful when tied to formal business processes, connected data and defined controls. Its latest platform changes are designed to make that model work for both business users running automated processes and developers updating applications.

The new functions cover agent interoperability, shared data context, developer tooling and cross-process memory, extending Appian's effort to make process models the framework through which AI systems access information and take action.