Magazine Store

50 Innovative Companies to Watch 2020

An Interview with Bayestree Intelligence Leadership: ‘Sainapse helps customers get the right answer on time 90% of the time across complexities and languages’


“Sainapse is definitely onto solving a serious gap in the customer support space.”

When a potential customer is still a prospect, their questions could range across alternatives in the competition space, price performance, model options, availability, service and support, reviews, and so on. By contrast, their post-purchase portfolio of questions change to complaints regarding performance, breakdowns, availability of spares and service engineers, billing issues, etc. A McKinsey study confirms that not getting the right answer on time was the single biggest dissatisfaction driver at over 40 percent.

Over the last three years and more, Boston Consulting Group, Pricewaterhouse, and Deloitte have come to confirm the same home truth through their independent research as well.

Founders of Bayestree Intelligence—Avijit Biswas, Girish Koppar, and Partha Pratim Ghosh—were excited about this opportunity and set out to bridge this enormous customer support expectation gap via their patent filed proprietary Machine Learning technology-based product Sainapse.

As the pandemic made contact centers move to agents’ homes and force companies to find savings in each function, the Sainapse team is in a great position to help customers deal with the challenges times have unfolded for them.

Going digital, despite best of efforts by Chief Digital Officers and Expert Consultants aided by expensive Data scientists have so far rarely moved beyond analytical and peripheral processes in large companies. Part of it is due the way enterprises have gone digital – which so far has largely been via ‘process automation’. Achieving higher efficiency via coding for increasing number of digital rules made the whole digital proposition unsustainable beyond a point. Sainapse promises to help enterprises break free from this balancing act between efficiency and sustainability, at least in the customer support space.

We caught up with Avijit Biswas, co-founder and CEO of Bayestree, during the no-fly times. Avijit was excited to narrate the story about how one of the largest Japanese groups, a global Fortune 50, is redefining its customer support across 14 languages using Sainapse.


  • Automatically converts incoming messages into cases/tickets in a CRM/Service Management data store with appropriate summary, classification, and routing. (Impact – no human triaging)
  • Learns about possible solutions from all sources of knowledge (that’s typically used or referenced to resolve such tickets) and assists service engineer/agent by recommending solutions. (Impact – search time reduction by >80% and sets up for autonomous)
  • Optionally, ingests solution recommendations directly to the requester/end-user. (impact – self-help and reduction of expensive contact centres)
  • Triggers remedial action via script/RPA and calling the appropriate transaction in enterprise systems such as SAP/Oracle/SF/Pega et-al. (Impact – instant resolution, higher NPS, lower ops cost)
  • Spots emerging themes before it becomes frequency heavy. (impact – opportunity to take proactive action and stop a Covid)
  • Identifies names and addresses from data and has the ability to blank them out. (impact – zero risk of PII getting exposed)

Net-net, you could think of Sainapse as ‘intelligence inside’ a CRM/SM system that powers up a customer support landscape. Sainapse is language-agnostic and can read from any attachment and Avijit claimed even engineering drawings, images, PDFs, and emoticons.

We asked how industry-specific Sainapse can be and how scalable its architecture is. Avijit asked for a minute and invited in Partha Ghosh, co-founder and Chief Scientist. As Partha was joining the call, Avijit described him as the ‘science behind Sainapse’. After a couple of minutes, Partha online now from Hyderabad, India walked me through the extensible, event-driven architecture at the heart of Sainapse. He mentioned, “Sainapse is capable of learning from non-stationary data which makes it truly enterprise and Telco Grade scale ready”.

We then moved to discuss the industry focus topic. Sainapse, Avijit explained, learns from enterprise data that it has been given access to. Sainapse reads and learns from CRM records of past years as well as all ‘knowledge sources’ that typically store product manuals, standard operating procedures, bill of material data, tips and tricks, forms and next-step suggestions for support agents– pretty much anything. Therefore, Avijit explained, Sainapse can be used across Pharma, Telco, Automobile, and Banking without any changes to the core product. Avijit spoke passionately about programs that Bayestree is running across multiple global engineering industries. Partha continued and said, “Sainapse is even supporting project engineers in the field as they look for answers to their engineering questions”. Avijit mentioned, “to Sainapse, anyone asking a question or needing help is a customer. All that Sainapse needs is to learn from the knowledge of how such a question was responded before and have access to knowledge that’s typically available to an employee. Sainapse does the rest. It finds the paragraph in a 300 pages manual or row in a 20 million record database if that’s where the answer is”.

Upstream in the customer support process, Sainapse can read messages from across any channel (digital, social, or voice) and understand what the customer is talking about without needing any human intervention. Sainapse uses its patent-filed IntentExtraction technology to do that. IntentExtraction also tells Sainapse which message needs a response and which one is junk. I could relate to that myself as I have seen multiple banks carrying staff only to read messages and group them into ‘no action’ as well as ‘meant for further processing’.

At this stage, I was curious to know how Sainapse itself plans to keep adapting to the never-ending changes in the digital world. Avijit again paused, asked me for time to pull in Girish Koppar, co-founder and COO of Bayestree. Avijit introduced Girish as someone who was part of “steering SAP out of the dot com ditch as a founding member of SAP Markets team” and now drives product management and customer success at Bayestree.

Girish was now online as he greeted and shared with me, “you know we are now going GA with Sainapse 1.9”. “Sainapse 1.9 takes us miles ahead of our cohort in terms of both native capabilities as well as extensibility. We have an unmatched data adaptation framework that allows north or south side integration with any data source of nearly any data type. In addition, the workflow definition framework and our Edge Services are designed exactly for what you are asking – to keep evolving with new enterprise processes. Customer Support process, running on Sainapse, can easily onboard new channels, languages or even leverage external capabilities while Sainapse continues to drive the core getting the right answer on time”

I was still grasping and noting stuff down as Avijit added “we will have Root Cause Analysis even for unseen problems in Sainapse 2.0”.

We were by then well over our planned time when I asked how Bayestree was taking Sainapse out to the market – especially in this pandemic. In response, Avijit mentioned how Bayestree was following the same mantra they had learned in the mid-90s at SAP – ‘go to market with system integrators, consulting firms, and technology partners’. He said that Sainapse delivers around 90 percent accuracy in 90 days from go live starting with 70 percent out of the box after initial training. “In less than six months, it becomes clear to our customers that they have opportunities to save 50 percent of existing support cost.” He added ‘with this kind of impact our Partners are as excited as we are to go out and solve our customers’ problems together’

He further added Sainapse 1.9 is launching with native capabilities to isolate names and addresses and blank them out which makes Sainapse algorithmically GDPR compliant. Partha is also planning to expand the core technology vision by steering Sainapse into the world of concept learning.

Sainapse is definitely onto solving a serious gap in the CRM space. Forrester’s March 2020 CX megatrends spoke about embedding AI to create the ‘Super-Agent’ who’s focussed on only personalized complex issues. Sainapse seems to have cracked that to a very large extent.

The Leaders at the helm of Bayestree Intelligence

Avijit Biswas, co-founder, serves as the Chief Executive Officer of Bayestree Intelligence.

Girish Koppar, co-founder, serves as the Chief Operating Officer of Bayestree Intelligence.

Partha Pratim Ghosh, co-founder, serves as the Chief Scientist of Bayestree Intelligence.

“Sainapse reads and learns from CRM records of past years as well as all ‘knowledge sources’ that typically store product manuals, standard operating procedures, tips and tricks, forms and next-step suggestions for support agents, bill of material data – pretty much anything.”