Cutting Costs with AI in Customer Onboarding
By implementing AI, we've simplified our complex customer onboarding process. Customers now just need to give a prompt, and the AI takes care of the rest. It's a win-win for everyone.
Nobody has time for lengthy case studies these days, I get it. I've saved the details for our calls or interviews later. Here's a quick read about what I worked on. Estimated reading time: 2-4 minutes.
Problem Brief
Having a product that caters to many verticals can be an excellent business idea, but it has it’s own woes when it comes to customer onboarding. You will never know what a customer really want from the suite. To overcome this, we had a dedicated concierge team that gets into call with each customer and sets up a dedicated solution running. But the solution is not scalable as business grow. And at the right time, chatGPT came to rescue. We used it to build an AI based Customer Onboarding Experience.
Research and Approach
Trained AI models can understand human inputs to an extent and give reports based on them. We needed the exact model to understand the business of our customers. All we needed to do was, train the model about our products and all it’s features first, then get the requirements from the customer - match them - hurray! Problem solved.
Starting Step - Getting to know the business
The starting point of this approach is - to know what the cusotmers business is upto. We took two approaches here, one is to get the company website and domain details, and fetch a basic bio of the customer. Second one was to directly ask the customer what they do on a day to day basis. We took both the responses and fed it to Chat GPT
Matching the requirements and business operations
Chat GPT analyses both the responses and gives us a set of operations that it thinks are performed in the company. We present it to the customer and get a confirmation for the same. We know that chatGPT ar prone to errors, so there’s always an option to change the company brief again or even choose the required categories the customer wants.
A few more interrogations and app configurations
Once the user picks the required operations / category that they want, we proceed to asking even more detailed questions. They can choose to skip or answer - which will help us configure the best app for each use case. Around 91% of the people agreed to answer more about their business as long as we help them setup the product.
Results and Metrics
We’ve seen results of accurate app assignment to users based on their requirements and they’re not overwhelmed about the larger list of apps in the product suite anymore. We clearly tell them what each app does and why we’re adding them. What’s the icing on the cake? We’ve already configured each app from the users response and they work straight out of the box. Future plans are - to use this response in customer support, promote newer applications to user and provide dedicated training to them on newer features.
12hrs
of calls and meetings saved every week
33%
Faster adoption to products and its features
Feedbacks & Future Scope
Leveraging OpenAI for customer onboarding presented a significant challenge, but the resulting implementation has proven to be highly rewarding. It has resulted in substantial time savings during customer calls, receiving accolades from our concierge team, who can now secure more deals within the same timeframe. Our plans include extending this implementation to our website, where a chatbot will be able to address visitor inquiries comprehensively. We also intend to integrate the chatbot with our support system. This initial AI implementation has opened doors for numerous future opportunities, and we are eagerly anticipating what lies ahead.