Your next booking, handled by a voice agent

Hey, Obaid here again. This week: what we're building, a real result, and one AI shift worth your attention.

The phone rings while your team is already helping someone. The caller wants to make a booking. That means checking availability, collecting their details, and updating the schedule.

A voice agent can carry that work through during the conversation. This week, we’re sharing how one live agent connects the phone to the business’s booking system.

Client Win: One voice agent, 185 incoming calls

At one business, our voice agent has recorded 185 incoming calls between August 24 and October 9. It connects directly to the team’s existing software, giving it the information and tools to act while the caller is still on the line.

When someone wants to book, the agent can check availability, collect the details, and put the booking into the system. When plans change, it can find the existing booking and complete a reschedule or cancellation.

One call this week shows the full workflow: the agent checked availability, created a new customer record, and completed the booking in the same conversation. Each step was confirmed in the system’s action log.

That gives the team another way to serve callers while staying focused on the work in front of them. For requests the agent completes, the conversation ends with the system already updated—saving the team from taking a message and entering the same details later.

Worth Knowing: The cost of routine AI work just dropped

On October 7, Anthropic released Claude Haiku 5.5, built for fast, repetitive work like customer support, summaries, and database queries. Anthropic says it costs around 75% less to run on average than its previous Haiku model.

Asana reported more than a 30% reduction in task-completion latency in early evaluations of its AI Teammates, covering tasks such as setting up projects and finding overdue work. These are early results, but they point to a practical benefit: less waiting on routine tasks.

Read Anthropic’s announcement.

Our takeaway: lower model costs make it worth revisiting the repetitive work your team handles every day. Start with one workflow—answering a routine question, updating a record, or processing a booking—and measure the cost per completed task. The savings depend on the full workflow, not just the model price.

Have routine calls that leave your team with booking changes or follow-up work? Bring that workflow to a call, and we’ll map the opportunity with you.

Until next time, — Obaid