Intercom

🚨 Are you missing out on a goldmine of customer support insights? 🚨 We talk a lot about data in customer support – but are we making the most of it? Take customer browsing behavior for example. It's a real-time map of customer intent, but how often is it overlooked in the rush to resolve tickets? Chat transcripts are a vault of unfiltered feedback, but does it get buried instead of analyzed? And your agents are on the frontlines every day, spotting patterns and pain points long before they become problems. But are we truly listening to their insights? And of course there’s CSAT scores – easy to measure, but could we be digging deeper to uncover the “why” behind the numbers? What do you think? What's most underutilized data source in customer support? Cast your vote below – and let us know in the comments if your answer isn't included!

How can you tell if an AI agent is 𝙩𝙧𝙪𝙡𝙮 smart? Watch how it handles information. The smartest AI agents don’t just regurgitate answers from one place – they seamlessly pull details from help centers, internal docs, PDFs, and even URLs to deliver the full picture (see an example from Fin below) The basic ones? They’re just playing copy-paste. When your AI agent can integrate multiple knowledge sources, it gives customers the kind of comprehensive answers that actually solve problems on the first try. The result? Happier customers, less back-and-forth, and a major win for your team’s efficiency (and sanity).

AI-first customer service isn’t magic – it’s management. Great AI support starts long before the first customer interaction. It’s all about building the right foundation. Here are three steps to help you hit the ground running, courtesy of Natalie Hurst and the support team at Nuuly: ⇥ Establish a process to review and improve responses. Nuuly’s team did this with a dedicated Slack channel to immediately flag mistakes and improve accuracy. ⇥ Build out public knowledge base articles. AI can’t provide great answers without reliable content to pull from. Consider adding a specialized role like a knowledge manager to optimize this. ⇥ Define clear boundaries for what AI should handle versus what requires human attention – clarity helps keep your support reps confident and your customers happy. The secret is simple: put in the effort, and the AI pays it back in spades. Check out the conversation below 👇

Ever wonder how AI learns to “speak human”? That’s the work of conversation designers – the magic that turns robotic scripts into natural dialogue. If you’ve ever received an AI response that felt “off,” or doesn't match a brand’s tone, you’ve seen the need for this role. And as interactions with AI become a bigger part of our day-to-day, specialized roles like Conversation Designers are only going to get more important. In this clip from The Ticket Podcast, Fred (our conversation designer) reveals his favorite parts of being the architect that builds the communication bridges between humans and AI. For a behind-the-scenes look at a day in the life of a Conversation Designer, you can listen to the full episode wherever you get your podcasts!

Big news for customer service and AI fanatics: The 𝘐𝘯𝘵𝘦𝘳𝘤𝘰𝘮 𝘌𝘹𝘱𝘦𝘳𝘵 𝘗𝘳𝘰𝘨𝘳𝘢𝘮 is here, and it’s basically the VIP lounge for people who actually know their way around support, AI, and the future of customer experience. Here’s the deal: ➤ Certification? Yep, you’ll get the badge that proves you know your stuff. ➤ Free testing account? Have at it, tinker to your heart’s content. ➤ Beta access? Who's here to wait? Get first dibs on what’s next. ➤ Swag? Oh, we’ve got swag (the kind you’ll actually want to wear). Plus, it’s your backstage pass to the roadmap and a spotlight to show off your skills. Don’t wait for the invite – join now and make your move! 👉 https://lnkd.in/grZDZCzC

Working with an AI Agent has become crucial for customer service teams, but 𝙘𝙝𝙤𝙤𝙨𝙞𝙣𝙜 the right solution has also become increasingly complex. So here's a tip for choosing the right AI agent: Fancy demos can be deceiving. Sure, an AI agent might shine in a controlled environment, but how does it handle those messy, real-world conversations? Real customers rarely ask questions in the exact same way, so you need to ensure your AI agent can handle different types of communication. Test it with the kind of questions your customers 𝘢𝘤𝘵𝘶𝘢𝘭𝘭𝘺 ask: ↳ Tough questions that require information from multiple sources ↳ Different phrasings of the same question ↳ Incomplete or fragmented queries ↳ Questions with typos ↳ Various levels of formality Because in the real world, customer conversations aren’t scripted – and neither is great support.

Customer service isn’t a sprint. It’s not a marathon, either. It’s not even a single sport. It’s a little bit of everything. One minute, you’re playing chess, thinking five steps ahead of a tricky problem. Then you’re on the high-dive, navigating risk with precision and grace. Sometimes it feels like dodgeball, fielding challenges from every direction. But one thing stays constant: empathy. The ability to show up for customers, no matter what events the day throws at you. So, in a recent episode of The Ticket Podcast, when Bobby asked John Rossman what the main event would be in the Customer Service Olympics – a Customer Empathy Decathlon was the clear choice. You can listen to the full episode wherever you get your podcasts!

✨ Brace yourselves: Follow-up season is upon us ✨ It seems like everyone is circling back, but let’s flip the script. Instead of inboxes full of "Any updates?" emails – how often do you get a solid follow-up after a customer support chat? Whether it’s a "We’ve got this covered" or a quick resolution check-in; follow-ups are where support can get personal. But not every brand sends them. Which raises the question: how important are they? Maybe they're the secret sauce to leaving customers feeling heard, helped, and happy. Maybe they're tedious and unnecessary. Maybe it depends on whether or not the issue is resolved. What do you think? 👇 Vote below, and leave your thoughts in the comments!

Your AI agent’s intro could be the reason customers come back – or don’t. The greeting message is more than just a “Hi, how can I help?” moment. It sets the tone for not only the specific interaction, but the overall customer experience. Some quick rules of thumb: ↳ Don’t skimp on the details – more information leads to better questions. More detailed input = better answers. Every time. ↳ Lead by example – let your customers know the types of questions your agent can handle. So instead of vague “I need help” messages, you get sharper, more specific questions that the AI can answer more thoroughly. ↳ Get proactive – use available context to tailor the opening line, setting the stage and shaping the interaction. Accurate anticipation shows your customers you’re paying attention. This way, that generic, “Hi! How can I help?” message becomes something closer to... “Hey there! I’m your AI support agent, here to answer questions like ‘How do I update my shipping address?’ or ‘What’s your return policy?’ Go ahead, ask me anything!” Here’s Angelo Livanos talking about how Lightspeed Commerce approaches the all important greeting – and improves customer adoption and engagement with their AI agent as a result.

PSA: Stephen Fry and Brian Eno’s chat about AI leaves on January 11. If it’s still on your list, now’s the time to watch.