February 22, 2019

Next Best Action but then what?

Tim Miner

In this post, we discuss how AI can help sellers determine next best actions

How AI can help sellers determine next best actions

Both traditional Business Intelligence systems and new artificial intelligence (AI) applications tout the ability to analyze existing customer data and provide insight to sellers. This insight ranges anywhere from identifying which clients have a higher propensity to buy, to identifying which products are best to offer at the moment.

The latter is a great application for AI systems, called Next Best Action (NBA), since it takes many factors into consideration, like purchase history and cash position - an especially strong focus for financial services. These predictive analytics systems promise a significant increase in revenue by generating NBA insights. However, none have answered the question of this fundamental problem: How does the data get into the hands of the busy salesperson who has a full schedule of client meetings and deadlines to close business?

This is an enterprise architecture problem that has plagued AI applications since they were first developed. Insights generated are tested and proven to have the potential to increase revenue. However, our enterprise sales people have one major complaint: there are already too many systems for them to visit to gather the data they need to be effective. They hate their CRM system and they hate large reports they have to sift through to get the nuggets of information to be effective. All this research takes away from what they do best, which is build relationships with clients by spending more time interacting with them.

"Data insights are not being provided at the right time"

As soon as yet another system gets installed, this time with Next Best Action (NBA) insights for their clients, IT professionals are stunned to find sales not singing their praises. These insights have potential for increased revenue, so why the poor reception? IT just created yet another system for sales to go to, it’s the end of the quarter and the sales people need to make quota.

Despite many companies investing over $18B building AI systems in 2017, results are not paying off as expected. One of the major reasons outlined by consulting company ZS is that data insights are not being provided at the right time. This “right” time is when sales are walking into a meeting or preparing for the day’s events, instead of once a month when massive reports appear that need individual research time.

Redwood City-based Tact.ai has found a way to solve this problem. Instead of using AI to cull through customer data to get insights, Tact.ai uses its smarts to take insights already generated by other systems, sending them to the sales person at just the right time and at just the right place. The “right” place is usually their phone or increasingly voice activated systems like Alexa for Business since they make driving easier. The ideal time for delivery would typically be right before a meeting, where those insights can be best used.

"The ideal time for delivery would typically be right before a meeting, where those insights can be best used."

There really is not an industry where new insights regarding Next Best Actions (NBA) are not being generated by AI and machine learning systems. For instance:

●     A Pharma industry rep can learn prescribing patterns of a doctor they are about to visit.

●     A wealth advisor can learn which asset classes their client is well suited for before they meet them.

●     A server sales rep can learn which products had the most problems adapting to a computing environment based on support tickets.

All these insights promise greater revenue by providing NBA insights to the sales person. But if we don’t spend time creating a system – like Tact.ai provides – one that sends these insights to sales at just the right time, they will not realize that potential.

This “last mile” aspect of data and workflow cannot be forgotten.

Tim Miner
Tim Miner
Director of Professional Services, Tact.ai

Read more like this