Tailwyndz
Making AI Contextual to the Process

Organizations far and wide are dabbling in the AI/ML space to remain competitive in the market, increase revenue, and reduce operating costs. There's a lot of appetite for AI/ML in organizations which was apparent in the McKinsey Global Institute's Note from the AI Frontier. According to this report, around 70 percent of companies will be adopting at least one of the [varied] types of AI technologies by 2030, and less than half of large companies may be using the full range of AI technologies across their organizations. However, organization-wide adoption of AI technology is becoming a huge issue, where there is a lack of trust in the outcomes furnished by the technology. "The user wants to use the AI/ML algorithm in the context of the process, needs to still be in control of decisions and needs to trust AI/ML" says Kapil Khanduja, CEO, Tailwyndz.

Khanduja, along with Balaji Rasappa, understood this issue during their 20- year long career in delivering AI/ ML technology and together founded Tailwyndz to cater to this need of the organizations. The company is focused on addressing the user-adoption of AI/ ML technologies and making contextual recommendations to the business process. Tailwyndz offers the Expedit AI technical platform for various applications in the Retail, Consumer Goods, and Manufacturing industries.

The platform has an API base that makes it service enabled. Organizations can subscribe to the services across the cloud and get the recommendations wherever they want. Clients can choose to have the platform integrated to their enterprise supply chain planning systems or directly leverage the UI layer or app layer. "We have a decisioning engine to evaluate what is the right recommendation model for the organization," says Rasappa, CTO/CPO, Tailwyndz. "The decisioning engine automatically takes that decision and presents the right information to the user." The platform can learn from the changes made on the recommendation by the user.

During the implementation phase, Tailwyndz tweaks its pre-built models based on the organization's needs and expectations.

The organization's goals also act as the input for the platform. Based on the models' initial client-specific configuration and tuning, the processes are automated using automated feature engineering. These will be used to refine the model further and build a final program that can be developed as an API for real-time inference.

The user wants to use the AI/ML algorithm in the context of the process

Tailwyndz has had the opportunity to alleviate its clients' skepticism when it comes to adapting the recommendations delivered by their platform. One such client was a spare parts manufacturer who called-in Tailwyndz when they were struggling with their forecasting accuracy. The manufacturer had a vast network of customers across geographical locations, but the orders fluctuated, and the forecast was not helping. The client used traditional analytical engines to forecast, and there was massive skepticism from the forecasters when Tailwyndz was employed for the task. Tailwyndz now had to deal with two problems: provide a better forecast model and get the forecasters to adapt their proposed solution. The company added external drivers to bring in more data to improve the algorithm's efficacy, which resulted in the forecast going beyond the promised number. To get the forecasters to adapt the system, Tailwyndz allowed the users to select the AI/ML forecast or retain their forecast and proved through a “blind test” of providing forecast for six months that the AI/ML forecasts had a higher accuracy.

Now, the company is riding on their platform's initial success and is getting ready to take their solutions to places. Soon, the platform will be cloud-agnostic and will be expanded further with more pointed solutions, the best of algorithms, and the best of processes in the supply chain area.