ElevenLabs says its in-house chat helpers answer 5,000 questions and open 150 tickets a month, a small model for any team buried in routine requests.
ElevenLabs, a company that makes AI voice tools, says its own staff now lean on chat-based AI helpers for daily chores. The company claims they save more than 1,200 hours of work every month. They answer more than 5,000 questions a month, and they also take action, such as opening tickets and building customer demos.
The helpers began as a simple way to pull up internal information in Slack, the workplace chat app. Examples include user retention data, the right documentation, or the history of a customer account. They then grew into helpers that act on requests. A teammate asks a question, reports a bug or asks for a demo, and the helper answers or gets to work.
The triage bot created 150 new tickets in the last month. ElevenLabs says that work used to be very manual across its support and engineering teams. Before it opens a ticket, the bot searches Slack for background and checks for duplicates. It also makes sure there is enough detail to diagnose the problem.
It then files the ticket in Linear, a task tracking tool, with a priority level. It sends the ticket to the right team using an ownership document that was uploaded to its knowledge base. Teammates can ask follow-up questions in the same thread.
The product-focused bot can build a full customer demo on request. From one message, it assembles a working AI agent set up for a prospect's use case. ElevenLabs says its helpers have built dozens of customized demos so far.
The company points to two lessons. First, answer quality depends on what the helper can see, so each one gets carefully chosen data. Second, the helpers must be reliable from the start, or people stop trusting and using them. Because they work in shared channels, mistakes are visible. So are successes, which the company says helped adoption.
The figures come from ElevenLabs itself, in a post on its own blog. The post does not explain how the 1,200 hours were counted. It also does not give costs or say how often the helpers get things wrong. Teams weighing a similar project would need to measure those things in their own setting.
If repeat questions and routine tickets eat your team's time, a chat-based AI helper is a small project worth testing, with clean data and a way to track mistakes.
Otto, a new customer-tracking app for field sales teams, lets reps speak their notes instead of typing them. Price and launch details are not yet known.
Read more: Otto launches a sales app you talk to instead of typingWorld models, AI that learns physics from video, are being built into home robots and driverless cars, and your own videos may help train them.
Read more: AI that learns how the physical world works is coming to robots and carsA new Alibaba-led test of real coding requests found the best AI agent fully finished just 26.4 percent of tasks, so human review still matters.
Read more: AI coding helpers fully finish only about 1 in 4 real tasks
