Akshat Jasra reveals why executives are not using their already existing data. In this age of endless supply of computing power, data, and real-time data supply, companies notice that not many are adding any tangible value to their bottom line. Based on clients Jasra has been working with, these three issues stand out.
The first issue is that the whole AI and Data machinery must be “Answer First” Centric. Measuring AI projects this way can have two significant benefits. First, it gives a clear orientation of where to go. AI teams can get lost on training models, trying new techniques, and increasing accuracy. A business orientation allows companies to measure AI projects from a financial perspective, and we can estimate the cost, the revenue, or impact it will have on the company with KPIs like ROI.
The methodology should focus on “How can a company quickly reduce working capital” rather than large Cloud and infrastructure projects focused on technology.
The second issue is that easy to use “What IF” UI is still not being offered. Rather than waiting months to deliver an end software, it is better to work on tight, short iterations that rapidly produce outcomes. After each iteration, companies should receive feedback from customers and users to further streamline improvements. After that, hypothesize why the user will not love the solution; usually, this is related to model accuracy or performance.
The final issue is the lack of the right people wired for Data and AI. Data Scientist is a pretty recent job, and very few Data Scientists can say they have more than 15–20 years of experience. Data scientists are a scarce resource, and it is difficult and expensive to get a good team. More importantly, getting to know algorithms is no more a skill. With AWS, GCP, and Azure offering end-to-end tools, Akshat Jasra says companies need people who can put and use statistics in the context of the problem they are trying to solve. There are teams with folks having Ph.D. And yet, all they know is the mathematics of the problem, which the Cloud can easily do on its own.
Learn more with Akshat Jasra for additional details.