Intercom has launched three new features built using OpenAI’s GPT-3.5 technology, the same model that powers ChatGPT. According to the company, more than one hundred customers are already using and testing the features within the Intercom Inbox and its Articles product in a limited Beta, with wider availability coming later this year.
The new features include Composer AI, Conversation Summarization, and Article generator AI.
“We believe the advances made in Large Language Models (LLM), spearheaded by OpenAI and that have manifested in ChatGPT, will be the most significant change the customer service industry has experienced in decades. It’s a massive step towards a world where customer service is automation-first, a future we've been working towards for a few years,” said Des Traynor, Co-founder and Chief Strategy Officer at Intercom.
The feature enables agents to write down shorthand customer responses based on which the AI generates a complete suggested message using the new Expand and Rephrase capabilities.
In addition, agents can automatically tweak the tone of the message into more casual and happy or more formal with dedicated buttons. This way response rate is higher while the quality of messages is improved.
Conversation summarizationTo make the customer interaction handover easier between agents, an AI-generated summary of all conversations appear within the Intercom Inbox. Described as providing a "level of automated summarization not possible before," this feature lets support agents pick up exactly where a previous agent left off - without the need to ask the customer to repeat any information.
Article generator AI
When writing help center content with Intercom’s Articles, authors can paste a summary of the content they want to explain and use the AI to generate a full article version, which they can then edit or update as required. This way, teams can produce more content in less time.
Commenting on the seismic shift that the launch of GPT technology caused, Traynor added:
“We quickly built and shipped many prototypes internally, ensuring we learned and iterated with each round of user feedback, and we've released our first wave of features based on GPT-3.5. We're excited to get this into the hands of our customers, but we’re just scratching the surface of what’s possible with this technology. We believe the capabilities won’t just be point solutions, but will create a technology that will sit across our entire customer service platform. We’re actively exploring what comes next.”