Prashant Bhuyan, CEO of Accrete.AI, sat down with Robert Phillips, Director of Listings for Nasdaq who proceeded to first ask,
"What about Accrete? What problems are you solving? What do you do?"
Investors are drowning in information and it is mentally exhausting to search through all of this information looking for insights that could move markets. The consequence is that investors are ending up making bias decisions that lead to underperformance, our mission at Accrete.AI is to help investors overcome these biases caused by information overload, make better decisions, and generate alpha in the digital age of markets. We accomplish this by building smart investment tools that are continuously reading and understanding huge volumes of information in a very specific context and extracting actionable insights buried in the linguistic nuances of natural language. Accrete pushes these insights to portfolio managers analysts and traders so they can spend less time searching for information and more time making well informed decisions. We launched our company in May 2017, today we have 27 people on the tea, we are based in lower manhattan, we have world class AI scientists, industry veterans, that have a lot of experience building alpha generating tools. We are really excited about some of the trends we are seeing.
"What can you tell us about Accrete's approach to Artificial Intelligence?"
Prior to launching Accrete we spent many years trying to automate cognitive tasks to help human investors solve decision problems in financial markets. The primary challenge we faced was in order to generate accurate results you had to remove bias from machines and create systems that were adaptive to changing contexts. Fundamentally all of these problems are rooted in the lack of relevant training data. We found we couldn't solve very general problems, but could solve very specific problems, and we did that be leveraging human expert networks together with artificial neural networks to create an approach that produced results that are more accurate, efficient and scalable that what humans or machines could achieve independently. A core principle of Accrete is this idea of the 'Glass Box' we don't believe fundamentally that a human will find the outputs of a complex system like a machine to be useful unless they can trust those outputs. We deliver out insights through what we call a 'Glass Box' that exposes some of the reasoning behind our outputs and provide users with reproducible research so they can extend our work and find new use cases for our applications.
"What more can you tell us about Accrete's products?"
We have four products in the market today and several more on the way, we have; Rumor Hound which continuously scours the web for actionable M&A rumors, Topic Deltas measures and monitoring linguistic nuance and sentiment shifts in earnings calls, Data Distillery analysis the supply chain and extracts intelligence looking for risk and opportunities, and Rational Exuberance parses confusing fed speak.
Rumor Hound is really relevant today because of the M&A environment, and one of the big problems analysts, traders and portfolio managers face is that they are struggling to keep ahead of some of these market moving M&A rumors. For example, most investors are getting information from traditional highly credible news sources like Reuters, CNBC, Bloomberg... but these sources are often missing a lot of the market moving M&A rumors that emanate from countless digital data sources including chat rooms, Social media, articles, blogs etc. Rumor Hound understands exactly what a M&A rumor looks like including all of its linguistic nuance and its continuously reading 10's and thousands of different digital data sources and when it finds a M&A rumor, it pushes those rumors to users along with the target company, the potential acquirer, credibility score based on Alexa rankings, social media influence, as well as recency scores. Accrete transcends 'Black box' outputs as we include the actual source URL where the M&A rumor was extracted from and the rumor snippet itself so that users can validate whether or not that rumor makes sense and is actionable. We have been modeling these insights against market behavior and we are finding predictive patterns. For example, in January 2018 Rumor Hound detected M&A rumor chatter on Juno ($JUNO) a bio-technology stock and that rumor activity increased in momentum over the next couple of days and we had an 80%+ forecast that the rumor would impact the stock price by 5-10%. We were actually shocked that 6 days later on January 22nd that Celgene Corporation ($CELG) announced that they were buying Juno. We heard a big hedge fund was actually short Juno stock, they didn't get ahead of this rumor and turned into a mini disaster within their portfolio. Rumor Hound is really useful as a disaster insurance tool for a lot of these big funds, as well as an idea generation tool, an awareness tool, as well as a straight out trading tool.
"Why not just trade these ideas yourself?"
We think there is a handful of really sophisticated investment firms that are working on a lot of these problems but we see a greater opportunity in democratizing some of these technologies from a terminal value perspective and capturing that opportunity as opposed to just trading these ourselves.
"On the road to startup it's important to have key strategic partners, who are your key strategic partners and what can you tell us?"
When we launched Accrete we launched with the strategic backing of IBM. IBM provides us with scalable secure computational infrastructure, as well as enterprise distribution. What's really interesting (and is kind of a bigger trend we are seeing) big companies like IBM and their companies and partners are spread so thin in that they don't really dig deep into solving domain specific problems. In other words, they are not looking into trying to identify actionable M&A rumors, but a lot of their customers are demanding artificially intelligent tools that solve real world problems. So, IBM will take our innovation as a distribution partner and acting as a reverse channel partner they will distribute our innovation to enterprises wrapped up inside of enterprise integration services, support, compliance, data security. It's great for us because it provides turnkey integration into larger enterprises and we help bridge this enterprise innovation gap.