Most organizations have begun to invest in AI to guide their sales representatives as it helps organizations stay adaptable to changing customer needs and evolving markets. AI guided selling usually takes the form of machine learning generated advice offered to reps on their CRM or other software. It’s primarily designed to help salespeople stay organized, prioritize leads, choose the customer most likely to buy for their next sales call, and so on. When its impact is fully realized, it gives salespeople more time to sell and information that they leverage to sell more effectively. I’ve seen solutions like this work, even quite well. With such great promise, companies are investing in and designing their own custom solutions. Many start-ups are developing solutions and makers of CRM software have widgets designed to address AI-guided selling. In my experience most companies are struggling to implement this approach.


Although most of these initiatives fail, an investment in AI is more critical now than ever. But why do they fail? After many conversations with business leaders about these failures, I’ve noticed a pattern: Many leaders point to absence of sufficiently robust data to fuel machine learning models. Without enough data, AI can’t make accurate predictions and if the system fails to make salespeople more insightful or effective, adoption plummets. Sales organizations find themselves in this situation for many reasons, from the inability to stitch together a comprehensive view of customers to transactional data that resides in siloed repositories, to poor CRM data or the inability to stitch the data together to enable sophisticated analysis. But it’s almost always more than data that limits the success of such efforts and solving these challenges are essential to driving the success of such initiatives. These include alignment across leadership, leveraging first line managers, picking the right models and fitting in with current processes.

  1. Align leadership and evolve the metrics. Alignment among leadership on the vision for how AI will enable sales is critical to long-term success. This is a journey that requires alignment around different priorities at each stage. At first, AI guided selling may be about getting salespeople to try and then adopt the solution, while over time it may evolve to recommendation acceptance rates and participation in the feedback loop, and so on. While revenue matters, in the end, success is about embedding the right habits in place. Revenue and customer success will follow adoption.
  2. Involve first line sales managers. The importance of first line sales managers can’t be overstated. They play several important roles in any sales organization including but not limited to managing sales representatives, customers and the business. In AI guided selling, if managers aren’t engaged early and often, from bringing the solution to life to ensuring that reps participate, adoption will lag. So thinking through how to get managers aligned with the objectives of such an initiative and familiar with the details of such an implementation is critical. And since managers play an active role in guiding how salespeople strategize and execute in their territories, driving this kind of behavior change across the sales organization requires the integration of AI in all rep-to-manager interactions.
  3. The model must fit the need and be effective. First, models must fit their environments. For example, when enabling an inside salesperson who has responsibility for a hundred or more accounts, you’ll need a model that provides the right information about the right customer at the right time. Arguably, it’s difficult for a sales person to know and stay in touch with all customers across multiple products so this functionally is useful. In contrast, when enabling a salesperson with few, large strategic accounts, you’ll need a model that can coordinate sales efforts across myriad activities and individuals. Secondly, models must appropriately balance accuracy and explainability. With the current state of AI, greater accuracy comes at the cost of sacrificing clarity about why it’s made certain decisions. (The more complex the calculation, the harder it is to explain to a user with any efficiency.) If a salesperson receives a recommendation with limited explanation as to why it’s being made, she’ll be less likely to trust it, which will harm adoption.
  4. Don’t let data insufficiency stop you. Yes, better data leads to more robust predictions but finding adequate data is a journey for most organizations and while it’s a worthwhile investment, it’s likely to take time. Data insufficiency, however, doesn’t have to mean you do not begin. There are many paths to addressing this challenge, including beginning with educated guesswork while a more comprehensive data journey is considered. For example, organizations can direct salespeople to call on customers who have not been called on in the recent past. Such simple heuristics will generate customer insights that will be useful for guided selling and lead to increased sophistication over time.
  5. Augment the customer engagement process. AI guided selling solutions must enrich and augment the current customer engagement process. If an AI guided selling initiative requires salespeople to change the customer engagement process to accommodate, adoption will be difficult. AI guided selling solutions must fit within the natural workflow of the sales representative.

The machine is no replacement for a salesperson but if used cleverly, AI can drive greater effectiveness for the sales organization. Like most AI projects, it’s not an easy goal to achieve, but with persistence, the hype can live up to reality.