Hey there,

Have you opened this newsletter for a five-minute break from the never-ending wave of AI breakthroughs? Well, you’re out of luck; we’ve just announced one of the most exciting products we’ve ever built – and it’s powered by OpenAI's GPT-4 and Intercom's proprietary machine learning technology.

At the end of January, we released a suite of AI-powered features within our Inbox, using ChatGPT to make life easier for support reps in a variety of ways. Ever since, you’ve been asking us, “Can ChatGPT answer my customers’ questions automatically?”

Finally, we can say yes. Meet Fin, an AI-powered chatbot that’s fine-tuned for customer support. It’s fast, it’s reliable, and we think it’s about to become an indispensable partner to your support team.

What’s so special about Fin? Let us give you the headlines:

Fin uses GPT technology to converse naturally, interpreting complex questions and applying logic to produce our most conversational bot experience so far.
Fin is designed to answer questions based solely on existing content in your help center, increasing accuracy and trustworthiness. To reinforce that sense of trust, it always links to its source articles so your customers can validate responses.
Fin’s a team player. There will inevitably be questions it can’t answer – in those cases, it passes the query to support reps to work their magic.
There’s virtually no setup required; just point it towards your knowledge base and let it get to work helping your customers.

As with any of our products, we’re not finished yet. We’ve managed to reduce “hallucinations” considerably, but they’re still a risk we’re working hard to mitigate. Fergal Reid, our Director of Machine Learning, sat down with Des Traynor, our Co-founder, to discuss what we’re still working on, as well as the exhilarating possibilities of our powerful new chatbot.

If you’re nodding your head at the thought of AI-powered support, we’d love you to get involved. Sign up here to join the waitlist to try it for yourself.

To inFinity and beyond,

What we've been doing

ruth o'brien

How to keep and nurture customer support talent

Excellent customer support is driven by an excellent customer support team. But once you have those talented teammates, how do you make sure they want to stay, grow, and progress with your company?

Here, Ruth O’Brien, Director of Customer Support at Intercom, dives into the three most important things to consider when it comes to keeping and developing great customer support talent:

Create a plan to develop and nurture customer service talent, finding a balance between helping teammates expand their skills and maintaining focus on customers
Define customer service success and progression paths (Take a look at our Customer Support competency matrix).
Bridge the product-support gap with embedded roles within the product team.

Liam geraghty

The GPT effect: A new era of customer experience

Since the arrival of ChatGPT, we’ve been working tirelessly to figure out what GPT technology means for the customer service industry. But last week, we looked a little further, to see what fascinating use cases are emerging in other sectors.

We hosted a fascinating conversation with Ethan Kurzweil, Intercom board member and Partner at Bessemer Venture Partners; Fergal Reid, our Director of Machine Learning; Krystal Hu, VC and Startups Reporter at Reuters; and Talia Goldberg, Partner at Bessemer Venture Partners.

Top takeaways:

For now, it looks like large language models (LLMs) will augment, rather than replace, human capabilities – making us more efficient and productive.
The price of LLMs is likely to drop over time – but for now, it’s a key consideration for businesses.
Rather than taking a “one and done” approach to these models, companies will likely use a mix that suits their specific needs.

fergal reid

Introducing Fin: Intercom's breakthrough AI chatbot, built on GPT-4

When we started experimenting with building a GPT-powered chatbot, we had a number of design goals in mind. We wanted to build a bot that could:

Converse naturally, using GPT technology.
Answer questions about your business, using information you control.
Reduce hallucinations and inaccurate responses to acceptable levels.
Require minimal configuration and setup.

We believe that we’ve achieved this. We’ve just announced that we’ve built an AI-powered customer service bot that has the benefits of GPT-4 and Intercom's proprietary machine learning technology, and is suitable for business needs.

Conversation starter: A chat about our new AI chatbot

des traynor

Conversation starter: A chat about our groundbreaking new AI chatbot

In this episode of the Inside Intercom podcast, Intercom Co-founder Des Traynor sat down with our Director of Machine Learning, Fergal Reid, to discuss our new AI chatbot, how we built it, what it does, and what the next steps look like for this remarkable breakthrough.

Here are the highlights:

Our new AI chatbot can converse naturally using the latest GPT-4 technology.
Fin consumes information from your existing knowledge base so you can control how it answers your customers' questions.
At Intercom, we believe that the future of support is a mix of bots and humans. Fin seamlessly passes questions it can't answer to human support.
We’ve reduced hallucinations by about 10x, building constraints that limit Fin to queries relating to your business.

Response Time: Vol. 2

We've just published the second instalment of Response Time, the series where we chat with customer service leaders to find out what they think about support, book recommendations, and whether or not a burger is a sandwich (yes, really!)

Find out the strangest thing a customer has said to Line Granseth, Head of Customer Service at Morganlevering AS...

What's new in customer service?

Record levels of consumer rage: The 2023 National Consumer Rage Study estimates that businesses are putting $887 billion in future revenue on the line because of mediocre complaint handling.
Rise of the machine... customers: Gartner estimates that by 2026, 20% of inbound customer service volume will come from nonhuman actors like virtual assistants or smart devices.
The benefits of concrete language: Using words that are more tangible and specific in customer conversations has been shown to not only improve customers' satisfaction, but also the amount they go on to spend with your company.