The future is now - Customers Support with software is real!
Just a few years ago the idea to utilize robots for Customer Support operations didn't sound very nice and was usually associated with lack of efficiency and inconvenience. Most of us had that experience of communicating with chatbots or voice IVR systems and not being able to receive needed information. Not a great experience...
Well, in the last 2 years things have changed drastically - mainly thanks to significant progress in the area of Natural Language Understanding (NLU) and Natural LanguageProcessing (NLP). New AI algorithms in combination with a modern computing power have opened a door for fundamentally new happenings in Robotic Process Automation. Chatbots and Voice recognition systems have been continuously getting smarter, and by the year of 2020 reached the point when such systems can effectively understand users' voice or text inputs related to a certain business domain and effectively support a conversation providing necessary information.
The biggest change is that the technology doesn't need to follow a strictly defined script or specific steps anymore. Creating such steps was a huge challenge in the past as it was nearly impossible to predict all possible behavioural patterns users would follow. Luckily, it's not needed anymore - instead of following predefined steps nowadays the Machine Learning (ML) technology is capable of understanding users' needs and self-creating a pattern to address those needs.
Majority of the modern NLPs/NLUs used in chatbots are capable of easily recognizing users' intents and classify entities with high level of accuracy. For example, it doesn't matter how a user makes introduction:
My name is Dave.
People call me Dave.
Call me Dav.
In all of these cases, the NLU system would be able to easily grasp that a user is introducing himself and the main information here is the actual user's name - Dave.
But improvements are not limited to NLU/NLP matters only. Significant progress was also achieved in the way how technology finds the most appropriate response once a user's intent is understood. Various Knowledge Base integration capabilities emerged over the last 2 years and armed bots with all internal knowledge companies have been creating for years. These days it could be just a matter of minutes to integrate a chatbot technology with an internal database and teach it everything you've been adding there for years.
The conversational technology got way "smarter" and that resulted in significantly improved user experience and exponential growth of chatbots utilization in the Customer Support industry within 2018 - 2020. Going further, it is estimated that more than 85% of incoming support inquiries in the form of text will be successfully handled by chatbots by the end of 2022.
Handling phone calls
Chatbots have become great! But how about dealing with phone calls?
No worries, things are much better on that side as well - thanks to noticeable progress in the Automatic Speech Recognition (ASR) domain. Voice recognition tech has also improved a lot in many languages allowing accurate real-time conversion of human voice to text and the other way around. And that opened up the door for another emerging application area - smart or AI-based phone systems which many companies have started adopting recently.
In simple terms, in recent years the technology has finally matured to the point when VoIP, Speech recognition, Natural Language Understanding and processing components can be combined in a single system capable of supporting conversation with humans by voice, and most importantly - such integrations are not going to cost a fortune anymore. So yes, bots now are capable of effectively serving your clients not only via text but by voice as well!
For the Customer Support industry it clearly means that such solutions will quickly replace traditional pre-programmed IVR systems and eventually human operators as well (at least when it comes to providing the 1st line of support).
In Rokitech we have been very excited about all latest AI, NLU, ASR developments as this application area was directly aligned with our line of business. We started experimenting with AI-based Customer Support systems in 2018 with the initial goal to create the Customers Support solution which would be able to effectively take over more than 50% of workload from human agents by assisting customers via voice and text communication channels.
We have spent more than 2 years experimenting with various products, setups, algorithms before at the end of 2019 we concluded that our systems reached the required quality/efficiency standards. That was the moment when we decided to move to Production utilization.
We quickly agreed and arranged trial implementations within 3 client companies. To their big surprise after 2 months of integration and configuration work with Roki Team, around 40% of their incoming customers' inquiries were taken over by the implemented solution. Then later after 4 months of improvement work this ratio on average improved to 60% of inquiries handled by technology, and 40% still required human assistance.
This was an impressive result and the further improvement work is still ongoing in order to achieve the ratio when 70% of incoming phone and website inquiries will be handled by our chatbots and voice systems. Rokitech Team is very confident we will be able to hit this goal no later than the end of Q3 in 2020. And once achieved, our "Smart Chatbot" and "Smart IVR" solutions will be available to all Rokitech clients.
Smart Telephone Assistant demo
Below are the sample recordings demonstrating how "Smart Callcenter bot" can handle incoming inquiry:
- Example 1 (handling a simple job opportunity inquiry)
- Example 2 (identifying a need, collecting contact information and setting an appointment).