Organizations have rapidly adopted artificial intelligence, but a stark divide is emerging: those who are embedding AI into the core of their operations, and those who are treating it as a standalone tool. According to a recent Harvard Business Review Analytic Services survey, only a small share of resondents say their organization has largely integrated AI into workflows.
Ultimately, when AI operates outside the flow of work, it lacks the context and connectivity needed to drive meaningful business outcomes. Appian innovations over the last quarter are designed to bridge this gap by continuing to make it safe and easy to embed AI directly into end-to-end processes.
Release highlights:
- Deploy a faster agent engine with streamlined evaluations, seamless interoperability via MCP, and federated data access for Appian and Snowflake.
- Automate dense data workflows using complex spreadsheet extraction and auto-suggested model updates to improve document extraction accuracy.
- Scale securely with improved AI governance and real-time connectivity via Kafka.
The new paradigm for AI interoperability and execution
Building custom connections for every third-party platform creates development overhead and maintenance burdens. Appian’s new Model Context Protocol (MCP) capabilities eliminate this friction and change how developers and AI interact.
MCP provides Instant, bi-directional connectivity for Appian and third-party AI agents.
Developer MCP. You can now bring your own AI tools to the table. By leveraging MCP, developers can use external AI coding assistants like Claude via a command-line interface (CLI) to directly and securely interact with the Appian environment, accelerating application development without sacrificing architectural control.
Bi-directional AI connectivity. Connect your Appian AI agents directly to external enterprise systems like GitHub, Snowflake, and Google Drive without writing custom integrations. You can also securely expose your data fabric, process models, and expression rules as discoverable tools for third-party external agents with the Appian MCP Server,. This allows external AI assistants to safely leverage Appian's underlying logic and robust security.
High-speed parallel execution. Scaling AI requires raw speed and reliability. Appian agents can now run multiple independent tool calls simultaneously. This parallel execution eliminates wait times, significantly reduces costs, and gets high-volume automations to the finish line faster.
Centralized evaluations. A new, centralized experience allows teams to manage and run test cases in bulk to instantly verify performance across critical scenarios and track overall accuracy metrics without needing to supervise individual runs.
Actionable process intelligence. Process diagrams in Process HQ now visually identify exactly who or what—human, RPA bot, AI agent, or integration—is executing tasks, helping you monitor AI performance and identify new automation opportunities with ease. Organizations with strict data locality requirements can now run Process HQ entirely within their own self-managed Kubernetes infrastructure.
Orchestration and integration: the true drivers of AI value
The HBR study reveals a critical roadblock: a lack of integration across systems is cited by 31% of organizations as a top challenge to embedding AI into workflows. Over the last three releases, Appian has focused on breaking down silos to make enterprise data actionable for both humans and AI.
Listen to Kafka topics, filter incoming events, and automatically trigger process models or write records.
Federated data access (Snowflake). The 26.6 release introduces native Snowflake support, unlocking instant access to massive datasets without requiring you to sync your data. More importantly, it equips your AI agents with advanced Snowflake Cortex AI and ML capabilities.
Real-time event orchestration. Appian now fully supports event-driven architecture through deep Apache Kafka integration. Using the new event consumer object, your applications can listen to Kafka topics, filter incoming events, and automatically trigger process models or write records in real time.
Automate dense data workflows. Unstructured and dense data can slow operations. In fact, while 78% of companies already use AI for document processing, 52% of staff time remains consumed by manual tasks. To solve this, Appian has introduced intelligent model improvements that automatically suggest updates, allowing models to learn from mistakes. The new Spreadsheet Extraction AI skill processes complex Microsoft Excel files natively to identify and extract data across multi-sheet reports.
Streamlined multi-app deployments. To manage modernized, interconnected systems, Appian introduced “releases”—a powerful deployment tool that allows teams to group related functionality and deploy multiple packages across multiple apps in a single action. It even includes built-in alerts to warn teams if objects are shared between multiple active releases, preventing deployment conflicts before they happen.
Orchestration and integration: the true drivers of AI value
Executives are eager to deploy agentic AI to lower operational costs and streamline operations, but governance is severely lagging behind ambition. While 92% of surveyed leaders agree that AI agents need rules-based guardrails to operate safely, only 48% have actually defined them. Consequently, inconsistent outputs and a lack of governance are cited as major roadblocks for agentic AI.
Appian's recent releases address this governance gap head-on, ensuring you never have to choose between leveraging cutting-edge AI and maintaining strict compliance.
Evaluate all generative AI inputs and outputs across your entire environment with centralized guardrails.
Centralized AI guardrails. The 26.6 release introduces a centralized AI Guardrails tab in the Admin Console. These protections automatically evaluate all generative AI inputs and outputs across your entire environment for severe risks like prompt injection, toxic content, and PII leakage.
Execution-level AI cost monitoring. To ensure scalable ROI, the Execute Gen AI Skill smart service now returns the exact number of AI actions consumed during each run. It also uses an optional Usage Groups parameter to map that usage directly to specific departments, enabling precise cost attribution.
Strict model controls. Organizations now have environment-level controls to disable specific model families entirely, ensuring compliance with strict regulatory mandates while keeping applications operational via approved alternative providers.
Trusted infrastructure & custom LLM gateways. You can now connect Appian directly to your organization's Microsoft Azure OpenAI accounts, ensuring your generative AI traffic stays strictly within your established security boundaries. For the strictest governance needs, administrators can also route every AI request through a custom, internal LLM gateway.
Accelerate your AI maturity curve
The next phase of AI maturity depends on moving beyond standalone use cases. By leveraging Appian's latest orchestration, integration, and governance capabilities, enterprise leaders can successfully embed AI directly into the core of how work gets done.
Read the 26.6 Release Notes or watch the Product Announcement Webinar.
