IBM Research-Ireland

How AI is Helping Sellers

Share this post:

At IBM Research-Ireland, we are exploring how AI can help all sellers make better, faster decisions and ensure that clients receive the best possible experience. We’re looking at what makes sellers successful and assessing every stage of a sales opportunity lifecycle where complex decisions are required. Joining forces with IBM’s Chief Analytics Office (CAO) and Chief Information Office (CIO), we delivered SCORE, the Smarter Cognitive Opportunity Recommendation Engine, a tool designed to help sellers increase revenue, commissions, and profits by increasing sales made by our business partners (BPs). In its first year of operation in 2018, SCORE has been a tremendous success, contributing $100M+ in additional revenue through the acceleration of IBM’s sales cycle. In fact, SCORE received a 2019 Stevie Gold Medal for “Improving Sales Win Rates with Cognitive Modeling and Process Automation.”

Lifecycle of a sales opportunity

Accelerating the sales cycle

Prior to SCORE’s implementation, opportunity lead passing was a manual process and decisions were based on limited seller knowledge and past experience. The previous BP recommendation tool used static business rules and was not able to capture seller experience and feedback in real-time.

To help identify where we could drive change, the CAO conducted a study which uncovered the following observations:

  • Opportunities passed within 48 hours closed with a 10-pt higher win rate than those which took longer to pass
  • Less than 25 percent of opportunities were being passed using the best practice of “warm transfer” (relevant information about the opportunity)

The study led to the conclusion that in order to make an impact, we needed to speed up passing leads to BPs, improve the BP recommendations, and automate opportunity passing.

Our research team in Dublin immediately began working on a machine learning model that learns based on historical data and the expertise of BPs, leveraging the relationships between BPs, clients, and sellers to recommend the best BP. These recommendations are then tailored and personalized to the seller, client, and product using inferred knowledge while providing transparency and explainability so that sellers have clear evidence as to why a BP is recommended; this embeds trust in the system.

As with most AI machine learning solutions, the model is only as powerful as its input data: in this case, vital information only the sellers themselves can know. To address this, SCORE leverages novel interactive AI techniques to incorporate user feedback loops allowing sellers to reinforce or discount inferred information, or even provide new information such as prior relationships between the BP and the client or expertise in a product. It applies machine learning, social network analysis, and optimization techniques to data to assess experience, performance, collaborations, and other attributes to rank the best BP for each opportunity.

Enhanced recommendations SCORE great results

Our innovation tackles three key issues missing from many machine learning solutions:

  • Merging of recommendation systems with prediction models. Most recommendation systems present users with options that match the user criteria. We go a step further by recommending options, but also predicting the impact that each option will have on win probability.
  • Explainability. By leveraging the explainable aspect of the machine learning model, we are able to show users which key features are influencing the recommendation, allowing them to make more informed decisions.
  • User influence. Many machine learning solutions are like a black box, where user feedback is limited to accept/reject. By showing users the key features influencing the decision, we also allow them to provide feedback on these individual features so they can influence and bring in their own professional domain expertise.

Thanks to SCORE’s enhanced recommendation solution, IBM has been able to implement automatic lead passing resulting in:

  • Accelerating the sales cycle with $100M+ in additional revenue in its first year of operation (2018)
  • An increase in IBM opportunities passed from 14 percent to 17 percent
  • A 50 percent reduction in the number of days to pass
  • A 5-pt improvement in the win rate on opportunities passed to BPs

Future enhancements planned for SCORE

The BP recommendation component is just one example of how AI is transforming sales: to extend expertise, free up time for higher value activities, and deliver a better experience for our customers. Our team is currently working on bringing further enhancements to SCORE, looking at every stage of a sales opportunity lifecycle where AI can help. Please watch our video “Lifecycles of a Sales Opportunity” below, and learn more about how we are using machine learning to identify new clients, find the best channel to serve an opportunity, and build a winning sales team. We hope to bring these components into SCORE over the coming year.


IBM Research-Ireland, CIO, and CAO team members working on SCORE (L-R): Eloise Mercier, Massimiliano Mattetti, Inge Vejsbjerg, Oznur Alkan, Alice Chang, Sanjmeet Abrol, Adrienne Loper, Rajesh Vadde, Naveen Bangaru Kumar, Ben Dubiel. (absent: Elizabeth Daly, Bei Chen, Rahul Nair, Claire Tian, Nathaniel Jesuran, Alan Zwiren)

Research Manager, IBM Research

More IBM Research-Ireland stories

Quantum computation center opens

IBM continues its push toward large-scale adoption of quantum computing with the opening of the first IBM Quantum Computation Center in Poughkeepsie, New York.

Continue reading

MIT-IBM Watson AI Lab Welcomes Inaugural Members

Two years in, and the MIT-IBM Watson AI Lab is now engaging with leading companies to advance AI research. Today, the Lab announced its new Membership Program with Boston Scientific, Nexplore, Refinitiv and Samsung as the first companies to join.

Continue reading

Adversarial Robustness 360 Toolbox v1.0: A Milestone in AI Security

IBM researchers published the first major release of the Adversarial Robustness 360 Toolbox (ART). Initially released in April 2018, ART is an open-source library for adversarial machine learning that provides researchers and developers with state-of-the-art tools to defend and verify AI models against adversarial attacks. ART addresses growing concerns about people’s trust in AI, specifically the security of AI in mission-critical applications.

Continue reading