Stonly Launches AI Agents That Execute Customer Service SOPs

Stonly has introduced Business Process Agents, AI systems designed to automate end-to-end customer service workflows by following the same structured knowledge and standard operating procedures that human support teams use.

The company announced the launch on September 17, 2026, positioning the technology as a solution to a persistent challenge in customer service automation: the gap between answering simple questions and resolving complex, multi-step support issues that depend on context, business rules, and sequential decision-making.

How Business Process Agents Work

Business Process Agents operate by following Stonly Guides—structured documents that define the steps, decision logic, required information, and actions behind a support process. These guides can be used interchangeably by human agents, customers in self-service scenarios, or AI systems executing autonomously.

The agents handle processes that require conditional logic: checking eligibility, troubleshooting technical issues, processing account changes, applying policy rules, and determining next steps based on what happens at each stage. They can request information from customers, make decisions according to predefined rules, trigger actions in connected systems, and continue until resolution or until human judgment becomes necessary.

When a process changes, teams update the underlying Stonly Guide once, and the modification immediately applies across all uses—whether by people or AI.

Eliminating Duplicate Process Documentation

According to Stonly founder and CEO Alexis Fogel, customer service organizations attempting to automate complex processes typically end up maintaining two versions of the same workflow: one for human agents and another translated into prompts, retrieval rules, and AI configuration.

"AI has gotten very good at answering questions, but customer service is not just a collection of questions and answers," Fogel said. "The most important support work follows processes: check this, ask that, apply this rule, take an action, and make a different decision depending on what happens."

Business Process Agents eliminate this duplication by allowing AI to execute the actual processes organizations have already documented for their human teams.

Flexible Deployment Models

The agents can operate in multiple configurations. They may resolve customer issues entirely on their own, handle portions of a process while escalating specific decisions to human agents for approval, or complete procedural work and information gathering before transferring cases to people with full context already assembled.

This flexibility allows organizations to automate as much of each process as appropriate without forcing every interaction into either fully manual or fully autonomous service.

Why it matters

Customer service leaders face mounting pressure to reduce costs by automating more than basic FAQ responses, but complex support processes have remained difficult to hand off to AI. The requirement to rebuild SOPs specifically for AI systems creates maintenance overhead, governance challenges, and risk. By enabling AI to follow existing human processes directly, Stonly's approach could lower the barrier to automating high-value support work that currently consumes significant agent time—without creating parallel systems that drift out of sync.

Stonly Business Process Agents are available as of the announcement date and can be deployed directly to customers or used alongside human agents. Details were first reported in a company press release.