I posted the following on Linkedin earlier today. Would love to hear more about how you are managing support knowledge and content in an AI world.
You’ve got to feed the AI beast with the knowledge it needs.
That’s one of my biggest realizations when implementing AI for customer service. 👆
No LLM (large language model) will have the unique information that’s relevant for your business and your customers. Only your team can create that.
Creating a culture of knowledge management is almost as important as the core AI technology itself. Here are 4 key lessons I’ve learned at @intercom:
1. AI is only as good as the information you feed it
What if I told you knowledge management can make or break your resolution rate? One of our customers, Zapiet, deployed our AI chatbot Fin. At first, it didn’t work as well as the support team hoped. They realized their knowledge base wasn't up to scratch. Instead of giving up, they gathered relevant knowledge and fine-tuned their content to better answer customer queries. Then, they deployed Fin again and quickly achieved a 40%+ resolution rate.
2. Write like a human, think like a machine
AI interprets knowledge differently than humans. Structure your help content in a way that humans can read and machines can understand. If you've created an FAQ document that a human can interpret and you've added simple yes/no answers, the machine won't interpret those answers in the same way that a human does. You've got to expand on what you mean when you say “yes” and what you mean when you say “no.”
3. Upcoming launch? Nourish AI with NEW knowledge
At Intercom, when we launch a new product, we want Fin to resolve 50% of the questions and issues that our customers have related to that product. That's a pretty ambitious goal, but to date, we've been able to achieve that for all of our major launches because we have a proven process for creating content that’s relevant for Fin. When you’re about to launch a new product or service, spend time creating content that will fuel AI to be as effective as possible.
4. Regularly water your content to help it thrive
Don't just develop content once and forget about it. To consistently make the most of your AI engine, audit your help content on a regular basis, keep it up to date, and make sure that it’s still being used by AI to resolve customer queries. The sky-high resolution rate and ROI (return of investment) will be worth it.
How is your customer service team creating a culture of knowledge management? What lessons have you learned?
![](https://uploads-eu-west-1.insided.com/intercom-en/attachment/2dc3c3a9-54d5-4c14-8d00-792785fbc836.png)