4 Steps to Ensure AI/ML has the Right Data to Learn

Artificial intelligence and machine learning (AI/ML) are growing in popularity for sales. Learn how data helps AI/ML be truly effective for sales organizations.

8 min read

Artificial intelligence (AI) may not yet be a go-to tool for salespeople, but it has evolved rapidly over the last few years from the theoretical to the buzzworthy. As the market races to deliver AI products targeted at sales users, it’s just a matter of time – a few years, if not a few months – before AI becomes a trusted part of the sales professional’s daily technology stack, giving useful advice and admonitions throughout the rep’s day.

Data is the Key Differentiator

But don’t expect AI to work out of the box – it’s not that kind of technology. In order for AI to work, it has to be exposed to data in order to learn how to make correct decisions now and to continue to make them in the future. Without that data, AI is helpless, and without exposure to pertinent data from your own company and third-party data sources, AI will be generic and offer little competitive advantage.

However, this internal view only is all to often the norm. This is a risky approach, as it captures how your firm might perform, but that does not mean that your firm is RIGHT. Bringing in additional data sources to the system helps ensure that you do not simply make your existing “worst” practices more efficient. 

Gathering the Right Data

But when it comes to sales, remember that AI doesn't need to be a bottomless data pit. What it needs is focus on the data that matters. AI for sales is what is known in scientific circles as “narrow AI,” meaning that it’s focused on a limited set of tasks and thus built around a limited set of data associated with those tasks.

“General AI” would be a system like the HAL 9000 from 2001: a Space Odyssey, possessing an ability to build decisions based on an all-encompassing data set, much as a human might. That would require an enormous amount of data which itself would pose massive management challenges. Limited AI simplifies the issue, drawing on data specifically associated with the tasks it’s in charge of.

Collecting data to help accelerate sales is nothing new – the recent trend has been to arm salespeople with predictive analytics so they can connect with potential customers in a more intimate and personal manner and close sales in record time. Like AI, predictive analytics are driven by data.

The AI Data Specifics

Some companies have taken note of how data is the fuel that primes the pump for these next-generation tools. For example, enterprises are using Xactly Insights, an advanced analytics offering based on 13 years of empirical incentive compensation data.

This allows customers to visualize what best-in-class sales compensation looks like, benchmark where they stand against their peers, and gain predictive tips on everything from how to create world-class programs to how to thwart sales rep attrition. The data that sales AI will need to succeed is slightly more expansive than SPM data, although SPM data will play a critical role in allowing AI to understand the factors that motivate salespeople to close bigger and better deals.

Other sets of data that need to be exposed to AI include the customer demographics of CRM, content usage data and training information stored within sales enablement, deal details and compensation and configuration information contained within CPQ, and pertinent data pulled from other sales-related applications.

Add in rich-third party data such as industry organization datasets, government economic indicators, and various inflationary measures for different geographies and a company (and its AI system) begins to have a rich, intelligence data set to learn, understand and make needed improvements from to improve rep engagement and revenue growth. This need for data places some important demands on any company:

  1. End the Need for Manual Processes: The need for AI should spell the end of any manually executed sales management activities. Spreadsheets, pads and Post-it notes will no longer work, since the data they contain is impossible to expose to AI. They’ll have to be phased out.
  2. Find the Best AI Software: Your sales operations, sales management and IT must determine the data needed to train and feed the AI solution. Sales ops and sales management must specify what applications generate that data, and IT will need to find software that replaces manual processes and integrates with AI.
  3. Help AI Learn: Everyone involved in the sales process must keep an eye on the system. The suggestions AI makes today that help win deals may not work in the future, and if the data going back into the system are lacking, AI won’t be able to change as customers and markets change. Just as the nature, tactics and tricks of the sales profession have changed over the years, sales AI is also likely to change. It’s up to humans, who have gifts of innovation, empathy and intuition that can’t be programmed into AI, to make sure that their sales AI is delivering suggestions that continue to improve and adapt.
  4. Develop an AI Strategy: Sales management, sales ops and IT must work together to develop an effective strategy to teach employees how to use AI properly. AI is a sales tool, not a replacement for sales talent. Salespeople must learn the best ways to use the insights it delivers in the context of a sales relationship. Buyers probably don’t want a fire hose-style stream of information, for example. If AI presents a salesperson with a wealth of content, the salesperson must know how to offer this to the customer in a way that makes sense for that particular deal.

AI can become a crucial tool for sales, or it can become a gimmick that gets in the way as much as it helps out. The outcome will be determined by the data businesses collect, how they integrate it into AI, and how much attention is paid by sales ops, sales management and IT to keep AI on target. From there, the sky’s the limit.