The Future of AI-Powered Slack Support

“Psst... have you caught wind of the GPT sensation sweeping the tech world? It's high time we explore its potential to level-up our Slack support game!” This is the introduction GPT4 recommended I say to start this article. But this is the perfect introduction not because of the quality of its writing, but to note that even this article is Chat GPT assisted. And this small interaction between human and AI is a pretty good introduction into how we see the Slack Support and broader support world adopting ChatGPT, LLMs, and all the AI advances in the next 5 years and beyond.

At least once a week, there is a new post on one of the communities I am a part of that sounds like this: “Hey everyone, we just build a tool that helps answer customer questions using Chat GPT. Who wants to help us beta test?”. Not sure why, but this seems to be the first use case that people think about when they bring together AI and Support. There is nothing wrong with this, and there are plenty of companies like Solvvy and Forethought doing this really well even before Chat GPT. So then how do we move past this use case and think of more interesting or non-obvious use cases.

A few weeks after the launch of Chat GPT, Intercom made an impressive announcement, introducing a suite of features built upon Chat GPT. These enhancements facilitated an easier agent experience without having to create a whole new chatbot. Among the updates was chat summarization, which allowed agents to catch up on customer conversations without reading through lengthy exchanges. Additionally, a collection of about six Composer features were unveiled, providing tools to modify tone, rephrase messages, adjust length, and generate articles from shorthand notes. These capabilities acted as a bionic arm for agents – while they could work without it, their efficiency increased significantly with these features at their disposal.

Are all of these Intercom features completely novel? Not really, in fact even Intercom was attempting to build a summarization feature way before Chat GPT. Grammarly has been detecting and suggesting changes to the tone since  2019, Google has been translating since 2006, and CNN have been showing AI based summaries since it acquired LENS in 2018. Many of these AI systems may even be better at those individual tasks.

What makes Chat GPT and other LLMs great is it democratizes access to amazing AI models. If you are a developer or startup founder, you can  build on top of models more powerful than any start-of-the-art Intercom, Zendesk, or Salesforce NLP model. There is nothing blocking you from getting feature parity and building the next Zendesk competitor. As these tools become more widespread, the advanced AI features we see today will become standard tools that every ticketing system must have, such as SLAs and reporting.

In 1997, The movie The Fifth Element saw flying cars. While we have been dreaming about flying cars since the 60s getting us to the office faster, 2020 made us question if we even need an office thanks to Zoom, Google Meet, and Teams. The point here is though we know the facts, we are really bad at imagining how technology will truly impact us in 20+ year range and what kinds of technologies we will even need. All this is to say, we are super excited to see where Chat GPT and LLMs will take us not only in the 5-10 year mark, but in next 20+ years.

We believe that over the next five years, support systems will adopt AI-powered features and techniques equally. These features will assist with answering customer inquiries, writing articles, summarizing requests, and helping agents work more efficiently. Due to the low barrier to entry, all systems will need to support a minimum set of features to compete in the market. What's even better is that these features will no longer be exclusive to companies like Zendesk, Salesforce, and ServiceNow; even the smallest startups will be able to implement them.

As we continue to expand the technology over 20+ years, I don’t attempt to guess which features and products teams will create - but we know that they will not be the Ivory Tower differentiators larger companies leveraged in the past. They will be easily implemented by all thanks to Chat GPT and other LLMs.

‍