By Nicolas Berthiaume, Sara Saiz Pozo and Tyler Zeeman
For organizations across the life sciences industry, developing a training model that can be delivered in a fast, scalable manner is key to sustaining a competitive advantage in the market. With competition becoming fiercer due to a fast-paced, continuously evolving industry, stakeholders need to find innovative ways to launch and commercialize new products faster without incurring higher costs in training staff.
This is where training enabled by artificial intelligence (AI) can make a big difference.
Novo Nordisk is a leading, long-standing global healthcare company whose main purpose is to drive change to defeat diabetes and other serious chronic diseases such as obesity and rare blood and endocrine disorders. They do so by pioneering scientific breakthroughs, expanding access to their products and innovating to prevent and ultimately cure diseases.
For each of these products, the company needs to make sure that their sales representatives are adequately trained regarding specific medical terminology, pharmacological effects and so forth. That requires an impeccable training of each one of their sales employees, via a highly standardized training program.
In addition, to put scale into perspective, Novo Nordisk employs about 47,800 people in 80 countries and markets its products in around 170 countries. With such a large organization, maintaining highly standardized training programs uniformly across teams can be quite a daunting endeavor.
Considering the challenges, Novo Nordisk pursued a faster, efficient and more precise approach via digital transformation of their traditional training models. This would not only enable access to training anytime, anywhere, but also provide each representative with personalized feedback on their performance.
As strong advocates of the digitalization of enterprise training, Novo Nordisk is one of the early adopters of online and blended learning. As the year 2019 came to an end, they decided to take their digital training initiatives one step further. They began to envision an “agile learning program” that allowed them to extract successful business experiences and the best frontline sales practices to ramp up medical representatives (MR) better, faster, and on a bigger scale. As an additional challenge, all of this had to be achieved without putting additional strain on their Commercial Capability Development team’s staff or budget.
By 2021, the Novo Nordisk Commercial Capability Development team adopted a group selling activity AI training model as one of the main new solutions for sales enablement. With this model, they leveraged AI not only to enhance MR capabilities, including product knowledge and soft skills, but also to effectively drive sales performance into actual business outcomes.
Due to its enormous potential to transform the way we work, and specifically the way we learn, AI is being increasingly used by companies to improve employee training programs.
It’s worth distinguishing between two distinct types of AI: generative and analytical, both of them with promising applications to improve training programs at several levels.
Generative AI was conceived to create new output, whether in the form of text, sound or images. This type of generative AI is especially helpful when it comes to generating training materials, as it enhances the efficiency of instructional design and improves the effectiveness of employee training programs.
By contrast, analytical AI is designed to analyze and interpret data to identify patterns and provide insights that enable training programs to be tailored to individual employees. This type of AI technology becomes especially useful in professional fields like finance and healthcare.
Again, one of the cornerstones of effective training is receiving timely and targeted feedback to drive improvement. In the case of Novo Nordisk, thanks to analytical AI, this feedback on performance could be automated, standardized and delivered without any kind of bias.
In view of that, Novo Nordisk co-developed with UMU an AI-powered sales training tool that would provide both a safe virtual environment and the necessary coaching guidelines and feedback for repeated, deliberate practice on a scale never before thought possible through the use of analytical AI.
This sales training tool was developed based on the premise that outstanding sales pitches and business presentations are composed of the same common elements. Accordingly, the training model used AI technologies such as few-shot learning and zero-shot learning, knowledge graph, natural language understanding, automatic speech recognition and computer vision to give structured, objective and standard feedback on six basic analytical dimensions (see Figure 1):
All feedback on practice received by employees would strictly target Novo Nordisk’s own industry standards and best practices. With only a minimum of information (namely, a presentation deck, accompanied by video examples of what Novo Nordisk considered to be a good presentation and a bad presentation), 10 fully customized and reliable AI sets were generated for self-service practice. The project required no large-scale manual annotation of data, and proved fully controllable in terms of resource expenditure, namely staff, time and budget.
Novo Nordisk reported that sales trainees who had practiced with this AI coach showed an increased interest in their training, while also being more likely to retain product information. In general, trainees made outstanding, quantifiable progress: Their discourse became more coherent, product descriptions were more accurate, and their overall demeanor appeared more natural and c o n fi d e n t .
Within the first year of its implementation, this new training model was used by more than 3,000 Novo Nordisk medical representatives together. These medical representatives submitted 84,647 exercises (averaging 25 exercises per medical rep), for a grand total of nearly 23,150 hours of practice time.
Data analysis conducted by Novo Nordisk also showed that learner scores were in direct correlation with the number of exercises they had submitted via AI practice. Those whose progress was most obvious increased their AI scores by 70% within only five attempts.
On average, those who used AI in their training saw an 8% improvement in performance over those who didn’t, and those who submitted five or more attempts outperformed those who submitted only one attempt by 10%. Finally, in comparison with similar training programs conducted during the same period of time using traditional modalities, Novo Nordisk’s agile learning program reduced costs by 52%.
As a senior commercial capability development director said, “In the context of the pandemic, using AI instead of in-person trainers for large-scale coaching and evaluation has brought costs down and allowed us to support more projects. Our training program has evolved rapidly to adapt to the current challenges, and it has made training itself more efficient as a whole.”
Novo Nordisk also conducted further data analysis to evaluate the validity of the AI-generated feedback. It observed that the AI’s smart scoring results matched the “average results” provided by individual human trainers at over 75%, while offering more consistent key metrics. And when judging high-scoring or low-scoring exercises, its matching accuracy would rise even higher.
This is testimony to the fact that, unlike human feedback, AI-generated feedback is based exclusively on objective data; it is unaffected by motivational, behavioral or cognitive factors and biases.
The Novo Nordisk case demonstrates that AI technology is revolutionizing the way sales training is conducted, making it more engaging, efficient and effective than ever before. With AI, sales coaches can create personalized training programs tailored to each individual’s knowledge and skills level. AI algorithms can assess accurately each employee’s knowledge gaps and provide them with the necessary resources and training they need to improve their skills.
Moreover, AI can provide real-time feedback and coaching during training sessions, helping employees to make the necessary improvements as they go. In addition, this type of training experience can also lead to better engagement and retention of the training material.
Nicolas Berthiaume, firstname.lastname@example.org, is global program director at UMU.
Sara Saiz Pozo, email@example.com, is European program director at UMU.
Tyler Zeeman, firstname.lastname@example.org, is sales director at UMU.