Today as automation is accelerating at a pace never imagined before. According to a recent report from the Harvard Institute for Applied Computational Science, Goldman Sachs used to employ more than 600 human traders back in the year 2000. The number today has astoundingly come down to just two, while the rest of the jobs are being galloped by automated trading systems managed by about 200 system engineers. This was just one of those examples that signifies the juggernauting trend of automation; one of which is Artificial Intelligence – AI. The trend of automation and AI has grappled numerous industry verticals, two of which are healthcare and financial sector being the most prominent.
The pre-requisite to that trend is to help organizations understand the significance and the potential AI can bring to their business. Kortical, a UK based AI firm has been working tirelessly to help enterprises understand and achieve the above. Incepted in the year 2016 in London, Kortical’s vision and mission to drive AI within organizations has been in two ways. Through its AI consulting solution, the firm helps businesses understand the potential of AI by unravelling practical use cases and by executing Proof of Concepts (POC) parallelly. This is because Kortical genuinely believes that every business should be AI enabled.
The second is through its flagship AI based platform which is, The KORE. According to Andy Gray, the KORE is a platform for AI to create AI. “We use AI to create AI and Machine Learning Solutions. Unlike our competition, we cut down majorly on overheads and extra expenditures in trying to identify the best fit solution of solving a specific problem.” Before arriving at KORE, Kortical has been relentless in conducting numerous brainstorming sessions to come up with AI POCs and this has been a greater enabler for organizations to help them find the right peddle in accelerating their project or any venture.
One of which is its Accelerated Intelligence programme. Through the Accelerated Intelligence programme, Kortical aims to help identify opportunities for organizations and its stakeholders on certain key questions that are relevant to their current business proposition. By brainstorming on questions such as “how do you generate revenue”, “where do you see the biggest opportunities are”, and “what data is currently available”. Such questions are intended to evoke and unravel numerous business opportunities for Kortical’s clients and its stakeholders.
The next step is to establish a business case which is checked for its viability. A realistic check on the organization’s existing processes, infrastructure is conducted and further the rate of new adoption is all looked upon to drive the best pilot use case. Once the use case is considered, a POC is created which demonstrates the viability of the use case that shows how actionable and effective results can be achieved from applying AI.
THE KORE Platform
THE KORE is Kortical’s flagship AI based automated Machine Learning platform that automates data-science thus allowing rapid deployment of enterprise level AI. This as signified by Andy, fast paces the model building/creation and optimization process from months to a matter of days. Based on three major driving factors that is – Prediction, Automation and Interrogation, KORE has been an affordable reality to many organizations which once thought the conception and access to the above three as unimaginable. Through Prediction, businesses while deploying AI can outscore their peers who use traditional methods easily by 10-20%. AI Automation brings out a profitable business by automating a wide range of activities, while Interrogation helps businesses understand why the AI is making those decisions which gives invaluable insights and confidence in the AI models.
The platform enables experts/data scientists to focus on engaging in feature engineering rather than doing trial and error based model selection. A very intuitive and robust web interface, allows faster deployment of trained models be well integrated into existing systems and processes. Explaining one of the use cases, Andy narrates an incident of a paper was submitted by the facebook AI research group consisting of 5 researchers over a long period of time running 1000’s of lines of code to deeply submit and solve the problem. Upon implementation, The KORE platform was able to get within 2.5% of their results within half-an-hour of the training without any extra inputs. “If you get any good data scientist working on top of that, then they can actually start to work on it, understand the output and analyze as to why such decisions were made by KORE. It is a very powerful tool to accelerate the delivery of AI,” claims Andy.
A case study to evidence the importance of The KORE is as follows – a team of six data scientists were working on and off for 2 years to get a good model that could read user reviews and automatically tag it with over 180 categories like love it, great product, great scenery, characters, etc. They plugged the data into the KORE and a few hours later had a model that was 46% better than their original AI model and even 4% better than when they had outsourced humans to classify the categories. Using the KORE they were able to have the AI model up and running in production within a few weeks. Andy explains, “There are a bunch of platforms that are attempting to assist the data science process but either simplify it too much so the results are not as good or don’t help with productionising AI and helping businesses use the AI models they create . The KORE actually automates a lot of the process, more than any of the platforms we come across. It is also focused on deployment allowing you to get up to the point where you are delivering value, thus helping you iterate quickly.”
With a step into the future through KORE, the Kortical plans to establish its foothold on its two major markets which is the Healthcare sector and the financial sector, apart from making inroads into the telecom sector. “For us the rate of adoption will be faster. We are taking enterprise customers along our accelerated intelligence journey, as the customers are getting smarter. The future will be more about the platform and solving such problems better, faster and cheaper than current models.”
