September 2026 Freshworks product update
Freddy AI moves deeper into the core of service work
Building the next iteration of AI service delivery takes time and a deliberate approach. It comes from steadily making AI more useful in real-world service environments, giving it the context to understand the work, the flexibility for admins to shape its behavior, and the reach to support employees wherever they need it. That is the direction behind our latest Freddy AI updates.
Over the last few months, we have made Freddy AI Agents easier to build, easier to shape to service environments, and available across more channels. We have also evolved Freddy AI Copilot into a more continuous source of context and guidance for teams. Together, we believe these updates give our AI a more active role in service delivery.
AI agents that are easier to build and ready for more work
One of the biggest barriers to automation is often the work required to build it. For example, admins might know what should happen when employees submit requests: collect the information, check policies, route approvals, update systems, and confirm it all worked. The process seems clear enough; turning all of that into a functioning workflow might not be.
New natural language workflows change that creation process. Inside Freddy AI Agent Studio, admins can build new Freddy AI workflows similar to how we just described. Type out the process, outcome, and actions taken, and AI translates that intent into executable workflow logic. This creates a faster, more intuitive path from admin idea to working AI skill. Once validated, simply add that new workflow to your Freddy AI Agents for an easy power-up.
The next step is putting those skills to work. The new Freddy AI Agent for Email brings agentic resolution into one of the most established employee support channels. It can carry the conversation across multiple replies, continually building on the context employees provide to help until the right resolution is reached.
Read also: Build the future of AI-first service
The Freddy AI Agent is also better equipped to handle the wider ranges of context of real-world service requests. With new capabilities to understand files and images more effectively, a simple picture may, in fact, save your employees a thousand words and reduce the back-and-forth DMs needed to resolve an issue.
On the Freddy AI Agent design backend, filtered search gives admins more control over the knowledge an AI agent can retrieve. By narrowing the sources they draw from, teams can keep responses grounded in the information most relevant to different request types.
Together, these updates make our Freddy AI Agents a more capable part of service delivery. With new abilities to support broader ranges of interactions and take on more service work, teams can rely on them in more moments across the service journey.
How McLaren, New Balance & more turn service performance into a competitive edge
An evolving Freddy AI Copilot
Freddy AI Copilot is also evolving to play a more continual role inside the service desk. Today, teams can call on Copilot to summarize tickets, suggest responses, and help with specific tasks as they come up. We’re building a new proactive experience on top of that foundation. The Copilot now surfaces useful insights and resources from knowledge and historical tickets as work unfolds (without waiting for an extra click or prompt). And with the new conversational interface, teams can ask questions, dig deeper into issues, and chat with their Copilot for guidance to the right resolution.
New agentic configuration also gives admins more control over that experience. Teams can provide clearer instructions for how their copilots should behave, what they should prioritize, and how it should reflect their service practices. Altogether, these changes are transforming the Freddy AI Copilot to become more contextual, more coach-like, and more useful in the flow of work.
Smaller improvements matter, too
Progress also depends on the details that make AI more useful and dependable every day. With MCP Gateway, Freshservice can connect more easily with external AI tools and third-party systems, giving AI access to more of the live service context and actions needed to get work done.
Behind the scenes, we have also been improving translation accuracy, ticket field suggestions, Writing Assistant behavior, and related change recommendations in the Copilot experience. Meanwhile, enhancements to Freddy AI Agent Studio make it easier to add new capabilities to AI agents thanks to custom functions, preset workflow triggers, and richer interactions inside agent workflows.
Each enhancement removes a little more friction from the workflow and expands what teams can do with AI. Together, they make Freddy AI easier to configure, more flexible, and more reliable.
Building for the AI service future
What excites us most is the broader shift these updates represent. Freddy AI is becoming more useful across the service experience, more adaptable to the way each team operates, and more capable of helping move work forward with less friction.
Each innovation step we take gives Freddy AI a more active role in service delivery and brings us closer to a future where AI can take on more work while teams stay in control and focused on better outcomes. That is the service operations future we plan on delivering.
*Certain features and innovations discussed are available to select customers under the Freshworks Early Access program. For more details on timing and availability, please refer to the Freshservice August and September release notes.
