Ajay Pillai

THE CHALLENGE
Improve the ROI for activity or campaigns.

Based on an extensive study with global users using design thinking, I understood the challenge for marketers: a lack of consolidation of historical data, a lack of understanding of customer preferences, and a lack of an instant feedback mechanism.

THE SOLUTION
Working with the data science team, I created an AI-based plugin to provide users with actionable intelligence and recommendations using an ensemble of machine learning algorithms. This included a propensity model to understand customer personas’ preference for certain tactics over others, a recommender system bringing the aptest combination of tactics based on the targeted persona, and forecasted the results for any scenario built. I also created a regression model to continuously forecast the campaign’s performance and the effects of any market-driven change, allowing a continuous benchmark against targets and a next action module to recommend actions needed to keep on track against set targets when data suggests a possibility of improvement. All of this was built on a simple, engaging, and visual UI, with an integrated warning system intimating when data indicates that results can be improved.

The purchase order (POs) for procuring goods or services from the vendors requires reconciliation to check if the goods and services were per agreement by providing supporting documentation. An independent team then audits this to ensure compliance with the guidelines of the business controls team. This was done for about 90,000 purchase orders annually and was time-consuming, so I was called to help solve this problem.

I did a deep dive into the process, tools, and systems being used to understand the need, the process, and where the inefficiencies lay. The process and the audits could not be eliminated as they were central to maintaining financial credibility, but there were non-value added that could be eliminated.

I built a system that pulled guidelines from a central system via an FTP and eliminated the need for separate documentation control. It prompted users, sent messages and reminders when reconciliations were needed, provided the required documentation based on the PO, and then uploaded documentation quickly. The application then analyzes the documents, photographs, and translations via AI applications and checks against guidelines, providing instant results. This eliminated the manual audit of the process and rework.

The HQ marketing operations team came to me with the requirement of creating a dashboard for funds transfer for inter-country and intra-country transactions, as there were significant delays and rework.

While the intent of the request was clear, I wanted to understand the reasons for the delays to help reduce and mitigate the issues.

I put on a design thinking workshop with users across marketing, finance, and accounting. The outcome of the workshop was clear: reduce manual work and rework due to incorrect data but handle specific requirements for each country. I simultaneously iterated with the users and UI team to get the front end designed, and the development team to build interfaces and the infrastructure.

My solution: An intelligent automation system that could reduce workload and improve accuracy. I used intelligent robotic automation and AI system-driven algorithms to feed the data from finance systems, eliminating errors. An approval and signoff process was created to ensure that the SODs were maintained along with messaging and reminders. We got this done in a record two months to handle transactions for over 150 countries with a transaction base of over 5,000 entries per month per country.