CRM data isn’t enough to make Artificial Intelligence work

 In Artificial Intelligence, Blog

The vision of AI for Sales is to transform how people engage with customers. However 90% of activities occur outside CRM, and even the data that is captured in CRM and marketing automation isn’t high quality enough to feed AI’s training models. Who doesn’t deal with wrangling sales reps to enter meetings reports or update contact roles? I’ve a big concern that AI for Sales is going to fall flat.

Data Struggles

Every sales and marketing team I’ve ever worked with struggles with data quality and quantity issues and in many companies sales and marketing leaders have thrown up their hands saying they’ll never really people able to solve the problem. In the words of Darian Shirazi, CEO of Radius “If we don’t solve this fundamental data foundation problem, AI will fail to transform business process. It is impossible to predict outcomes or even make great decisions off predictions if the underlying data training the models is stale, inaccurate, or incomplete. Driving insight means we need to truly have a sense of our underlying data and tie those truths to our business decisions.”

There are many rule-based, and standalone systems out there that are starting to add AI to their marketing positioning. However, a rule-based system, that is maintained by a “middle-man” can’t keep up with the production and use of content and activity across the enterprise, and making inferences from a small set of data leads to a high false positive rate.

A Better Way

With all the focus on making applications smarter and trying to capture more data in order to make AI work, I believe we are missing a better way to do things, using all the places where reps already work and where the corpus of data in the enterprise already exists….. It’s right under our eyes!

Think about all the other systems reps actually use in their jobs: sales hubs, learning systems, portals, running meetings, presentations, sending emails, attachments, having phone calls, follow ups, more meetings….THOUSANDS a day. And all these live in different places: email servers, laptops, drives, online drives, portals, CMS, learning systems, KM systems, departments, teams….all SILOS. Content silos distributed across the enterprise. This is a massive amount of data and it actually looks like this…

Imagine if we could harness it?

The nodes in the graphic above consists of the sum, correlations, and derivatives of data associated sales rep activity, and not just what’s happening in one isolated system. It’s hundreds of thousands of relationships between, people, information, time, location, organizational structure, outcomes, sharing, email, revenue, value, and actions. This is data beyond what you can ever buy from a 3rd party and greater than anything that can ever be recorded in a CMS or CRM. It’s the lifeblood of what’s going on in your company.

Bottom line: we are living in a different era, but sales leaders still struggle with the same old questions around pipeline, cloning top reps and predictability. You are asked to grow revenue but you have no scientific way to point you where to start. What’s really needed is a way to connect the dots between reps and activity. But this has been impossible up until now. To clone reps you have to know what they are doing across everywhere they work. Until then the promise of AI making salespeople better at their jobs is going to fall flat.

A New Way

The current buzz around AI for sales can only be sustained and achieved if the underlying data is complete, accurate and fresh. It needs to leverage feedback loops, network effects, the corpus of enterprise data (structured and unstructured), and be able to make inferences. To really allow AI to make effective recommendations a system needs to be able to make inferences about connected nodes and it needs to be able to highlight connections from derived data, correlation, and modeling. This is what we are delivering to our customers at Docurated.


Related Resources

Recent Posts

Start typing and press Enter to search