Bitcoin ATM network Localcoin: expert AI bots for instant, accurate support in Zendesk

• When you’re Canada’s biggest Bitcoin ATM network, you get a lot of customer questions from people who want answers quickly. For Localcoin, that meant a ton of complex, time-sensitive support tickets were piling up.
• They built a team of AI bots with eesel to get things under control. By connecting eesel to their Zendesk, Intercom, and internal docs, they created one central brain for their support.
• It was a seamless transition to adding eesel AI to their existing Zendesk messaging widget – no fuss and no data or workflow transformation necessary.
• They used custom AI Actions that could automatically look up transaction data – a task that used to be a huge time-sink for their human agents.

Who’s Localcoin?

If you’ve used a Bitcoin ATM in Canada, there’s a good chance it was one of Localcoin’s. They run a network of over 950 machines and serve more than 140,000 customers. People use them to buy and sell crypto through their ATMs, online, or with their wallet app. Crypto moves fast, 24/7, and customers expect support that can keep up.

Crypto moves quickly – so Localcoin needs to keep up

As Localcoin got bigger, their support team started feeling the strain. In a heavily regulated and constantly evolving space, customer queries were becoming harder to handle.

A lot of these tickets were tricky and urgent. An agent would have to manually dig through backend systems just to find a transaction status. This ate up time, frustrated customers, and created a serious logjam. Sorting and sending tickets to the right person was another manual headache that slowed everything down. Operating in a few different countries and a heavily regulated space, Localcoin knew they needed a way to give fast, accurate, and consistent support around the clock.

Expert bots for accurate, tailored answers

Localcoin didn’t just plug in a single chatbot and call it a day. They used eesel AI to create a whole system of bots working together.

First, they gave the AI a brain

They started by creating a single source of truth. They hooked eesel up to their Zendesk (Help Center, old tickets, macros – the works), Intercom, their website, and even internal Google Drive folders and PDFs. With all this knowledge in one place, the AI could give the same correct answer no matter where a customer asked.

Then, they gave each bot a specific job

Instead of a one-size-fits-all approach, they set up a few different bots for different tasks:

• The friendly front-desk bot: This is the first point of contact. It figures out where a user is from and hands them off to a specialized bot for Canada, Australia, New Zealand, or Hong Kong. No more generic answers.
• The super-organized internal sorter: They have an AI agent working behind the scenes, too. It reads new support tickets, figures out what they’re about, and gets them ready for the right human agent to take over.
• The walking, talking compliance manual: The team also built an internal Q&A bot trained on all their compliance documents. Now employees can get instant, verified answers to dense regulatory questions, which cuts down on risk and training time.