Bayestree Intelligence: Bringing Intelligence to B2B Customer Support
"I don't get my answer first time right," and "I don't get my answer quickly." These are the most common customer complaints in the B2B customer service space. A statistic from before the pandemic was that 40 percent customer would leave a supplier if they didn’t get answer first time right and don’t get it immediately. Covid crisis has only added rocket fuel to this - with more than 50 percent of enterprise staff working from home led from the front by some of the biggest tech names.
“Solving a CX problem requires powerful arsenal to deal with heterogeneous, semistructured data often with unknown schema. Bayestree's new innovation called Information channel (IC) comes handy here. It acts as a signal conditioner ensuring standard outfit for heterogenous data. Post conditioning, underlying inference is extracted using Sainapse core ML capability. Inference is now ready to be consumed for significant business gain -Partha Pratim Ghosh, Co-Founder & Chief Scientist”
When B2B customers report a problem, especially about a high-value product, chances are that their operations and revenue flow is halted or at the minimum impacted by the down time. The product could be a critical component in their assembly line or a generic drug required for creating a vaccine. Often times getting timely responses holds the key. However, most often, that does not happen. Root of this issue lies in knowledge scatter and widely varying competence level of support staff – both inherent in most corporations. CRM or other enterprise systems that engineers and agents sift through to answer a support query usually do not provide get to what one if looking for, for example, one of the most frequent calls after buying a car is around 'how to save an FM channel as favorite'. Probably, the answer lies on page 37 of the manual, and the agent ideally needs a system that takes her there. Or still better – the car buyer can get to Page 37 on her in-built infotainment system just by asking and avoid the agent conversation altogether. But that does not quite happen.
Another primary setback agents' face while solving a customer query is going through and understanding various data types. Customers of high value equipment often report details of their problems via email or company provided support channel and attach images, associated drawings and of course description in text. Imagine how long it would take even for an engineer to sift thru this, make sense and start problem solving – when production line of the industrial customer, dedicated bandwidth for a software company or a path lab of a clinic is hit.
Based on the Bayesian technique, Bayestree Intelligence provides the much-needed 'intelligence' to service management systems like CRM and assists engineers and agents alike in offering first time right resolutions recommendations to even complex problems so that product line or path lab can get back to work. Of course the car buyer who asked for ‘how to save a radio station as favorite’ got her answer without bothering the agents at all.
The company's AI product, Sainapse integrates with and works behind any CRM system—be it from SAP, Salesforce, Oracle, or Microsoft—to address B2B customer service space issues. Sainapse understands customer problems in full context – reading attachments from XL files, PDFs to even engineering drawings and is not limited by set words, text mining rules, or error codes. Best part is Sainapse extracts intent behind the question and processes based on that understanding much like humans would do in real life. So asking Sainapse a complex question about a newly launched product in an automobile company would take the caller to appropriate engineers while asking about Superbowl score in the same automobile company context would encourage a response like ‘Hey I don’t track that but hope your side wins.’
Sainapse uses our custom-made algorithms designed specifically and only to solve customer support problems, unlike platforms offering generic algorithms that could be used for preventive maintenance or building pricing models or even a game
"At a minimum Sainapse drops resolution time by 80% even for complex customer problems," says Avijit Biswas, CEO of Bayestree Intelligence. Biswas compares Sainapse's intelligence to that of a human. He says ‘treat Sainapse like a new joiner and giver her access to the same set of product manuals, design documents, databases, systems, anything else that you would share with a human engineer. Sianapse will read through all of them and learn much like you and I do’. Interestingly, while a human can read one page within 7-8 minutes on an average, Sainapse can read two million records or lines in under 18 minutes.
Built for next-gen customer service, Sainapse can easily be implemented by clients without making significant process changes or ripping through existing tech investments. When a customer query comes in, Sainapse quickly identifies incoming language and switches to an appropriate knowledge base for reference. Sainapse can comprehend a combination of languages as well, like English and French or English and German. "Sainapse is the only AI product that can understand all global languages and operate with all languages simultaneously," says Biswas. Besides, the product's attachment reader reads through images and screenshots to understand the context better. This way, Sainapse claims to eliminate think and search time in problem solving.
Sainapse's multi-dimension classifier identifies the appropriate topic, sub-topic, priority, assigned group, and any number of critical classifiers that are needed. Businesses need not worry about duplication as Sainapse can check for potential duplicates and then creates a case in the underlying datastore. The solution can escalate cases that are new and not covered. Once the claim is created, Sainapse recommends remediation options by sifting through numerous knowledge sources, including multiple instances or different data stores, to close the loop with customer acceptance information.
What is more, Sainapse brings the ability to spot the emergence of potential problems and report them before they turn into a crunch. Biswas points out that customer problem solving usually is about staying alive to situations that may be getting reported for the first time. Customer service teams depend on the frequency of occurrence to prioritize focus and budgets. Unfortunately, by the time an issue makes a billboard list, it has created significant customer side challenges and frustration. And there lies the real problem.
Sainapse Theme Discovery Engine keeps building holistic understanding of customer queries and reported issues to spot budding themes that could even be a black swan. Biswas says ‘using modified variational inferencing techniques we are spotting next quarter's burning issues today, allowing agents to study identified data clusters, check adjacencies, attach an appropriate name and start problem-solving weeks or even months ahead of time’.
Sainapse offers Insights for next step actions by giving agents access to volumetric, remediation logs, recommendation metrics, performance measures to build dashboards that can be embedded in the customer landscape.
Bayestree Intelligence has entirely built Sainapse with its patent-pending algorithms. "Sainapse uses our custom-made algorithms designed specifically and only to solve customer support problems, unlike platforms offering generic algorithms that could be used for preventive maintenance or building pricing models or even a game.
Excelling Customer Support during the Pandemic
Elaborating on the current COVID-19 situation, where work from home has become the norm, Biswas says – “Maintaining customer experience levels has now become even harder. Support staff working from home, severely shrunk Field engineering capacity, sudden disappearance of hands on service engineer training are challenges that all businesses are grappling with. This is where, Sainapse’ ability to fetch the right answer from the right page and open up the right image that shows ‘how to fix’ become so much more impactful.”
In the current situation, companies using Sainapse can ask their customers to send messages via their dedicated support channel and receive instant response with detailed procedure to fix the problem. If Sainapse feels that its not trained enough to recommend a resolution to a customer’s query she would then get instantly escalated to appropriate expertise group without being forced onto either long wait times or music. With this change, businesses can instantly see that 50-80 percent of issues are solved remotely.
Few months back engineers from drive manufacturing units would typically visit a steel plant and fix the problems listed on the CRM service. Now, internal maintenance engineers of the steel company could act on most problems aided directly by Sainapse. And in case the drive doesn’t come back to life a message from the steel plant engineer who had tried out all that Sainapse earlier recommended, would take the case straight to OEM’s Field Service Module so that someone can now visit and fix it.
This new way of working is opening up opportunities of taking out 30-50 percent of support costs while improving C-sat.
A large Fortune 100 Telco is using Sainapse to not just ride over the current crisis but fundamentally transform their global customer experience. By autosorting and triaging incoming messages in 13 languages in a completely rule free manner, Sainapse has released significant cost and capacity for the Telco. In addition, end customers of the Telco, operating across East Europe, Asia, Far East and Africa, are now experiencing a response time never seen earlier – Covid-not-withstanding.
Raising the Customer Service Bar
Helmed by Biswas, who has advised and worked with enterprise CIOs and providers globally on 'world-class' IT design and execution across his stints with SAP, Siemens, and McKinsey & Co, Bayestree Intelligence is all set to take the next step in the growth ladder. In the months, the company plans to enhance Sainapse's potential by adding capabilities that would make root cause analyses and responding to unseen problems an enterprise AI reality. Biswas talks passionately about adding capabilities to search for images so that the steel plant engineer could just focus her camera on the drive for Sainapse installed on drive manufacturer’s side can start diagnostics and recommendation process already. When we were closing, Girish Koppar, Co-Founder and COO of Bayestree had by then joined our call and quipped “By the way your privacy is never compromised by Sainapse as it knows how to spot names and addresses and blank them out.”
U.S. Customer Relationship Management market size is anticipated to register a growth of 14.0 percent from 2019 to 2025 as per a report by The Business Insights. And Bayestree Intelligence has the lead in this growing CRM market. Besides, the company plans to expand its geographical influence beyond the U.S. and India and has started working with customers in the DACH – Germany, Austria and Switzerland region. Bayestree Intelligence continues to work tirelessly to improve Sainapse and help its clients provide their customers with the best, efficient, and quick-response customer support they deserve. As Biswas says, "We believe we can help the world a little bit in fighting the current situation, especially managing costs and not just keeping the customer service bar where it is but even raise it by a few notches."