At DVELP, we use Google's Dialogflow AI to support your contact centre and help your business operate more efficiently, improving customer experience and saving costs as a result.
There’s a lot of media hype around AI, so it’s best to start this topic with what AI isn’t: it is not an attempt at recreating human intelligence. The clumsy or gimmicky conversation bots you’re familiar with aren’t AI. Contact centre AI is not always about fully automating your customer service function.
AI is machine learning. It allows computers to deal with simple and well-defined tasks, leaving humans to focus on complex human problems.
You shouldn’t have to talk to a human to check whether a shop has your item in stock. You shouldn’t have to talk to a human to cancel your debit card and order a new one. You shouldn’t have to talk to a human to change your registered home address.
These are opportunities for automation and we use AI to achieve that. We introduce Natural Language Processing (NLP) to your contact centre calls so that your system can understand what your customers are saying. This is the AI piece. Using Google’s Dialogflow, we can tell your system that if a customer says “I want to speak to an agent” this means that they want to be transferred to a human. Dialogflow will then automatically widen the understanding to different turns of phrase, like “Transfer me to a human”, “I want to talk to someone” or “Speak to person”.
AI means that your solution can make inferences and thereby get a good grasp of human speech. From there, we can ask the system to do things with that information. On our M&S project, our solution is able to pick out intentions from the customers’ phrases and use that information to decide the next step. Primarily, this is about routing customers to the right end point. Instead of cascading decision trees, you get a one-step route directly to where you want to end up.
DVELP’s M&S solution is also able to handle some customer interactions in their entirety. For example, if a customer were to say “I want 3 apples delivered to my home address”, the solution would be able to pick out the order details.
Of course, sometimes queries do need to be sent to an agent. For example, a query like “I want to make a complaint” should be dealt with by a human. Sometimes you need the human touch for empathy, or to provide a customised service for complex queries.
When an agent is talking to a customer, AI can play a supporting role. In our Airline product, we created a feature called Agent Assist. It means that when an agent is speaking to a customer, the solution will understand the speech in real time and suggest possible replies based on its knowledge of the system and of previous customer interactions.
The future we envision for contact centres is one where the agent doesn’t have to use their keyboard or mouse. They can purely focus on talking to the customer while the system understands the speech, processes the data in real time, prompts the agent with best solutions and automatically updates the clients records.
This would mean less frustrated customer service representatives and lower staff turnover. It would mean quicker response times, more efficient processes and higher customer engagement. It would mean cost savings that would open the door to new opportunities of value creation in your contact centre. All of it would be made possible with AI.
The world of contact centres is changing. Speak to us if you want to know how.