Using AI in workforce learning creates business value, and most don’t see it.

mediumThis post was originally published by Jon Lexa at Medium [AI]

Image for post
Source: Sana Labs

In a report by MIT Sloan Management Review, BCG Gamma, and BCG Henderson Institute, seven out of 10 companies reported minimal or no return on investment (ROI) from their AI initiatives, while nine out of 10 survey respondents agreed that AI represents a business opportunity for their company. How can this gap between ‘perceived’ and ‘realized’ value happen? The simple answer is that somewhere along the roadmap from when an executive decides that AI is the right investment to when the initiative is executed, detailed plans fall apart — either the strategy was misconceived or the implementation was not well-planned.

In this post, I will share decision criteria for why and how Chief Learning Officers (CLO) should think about investing in AI, and then dive into how AI can transform a CLO’s learning agenda to bring real economic value to their business (e.g., an increase in revenue, saving time).

Decision criteria: To use or not to use AI

Using AI is not about adopting the latest technology. It’s a symphonic assembly of talent, process, culture, and strategy. When it’s orchestrated effectively, it’s beautiful. When it falls apart, it’s messy and expensive.

The first step is to think about the challenges your organization faces and the outcomes you want to achieve. Then, evaluate the potential investments you can make to create successful solutions. Sometimes the thorniest of business problems can be solved with a few months of simple programming, a few change management workshops, or diligent project management — involving no sophisticated machine learning algorithms.

The questions in the AI Investment Decision Criteria below serve as an important litmus test for whether or not your company should invest in AI to solve a business challenge and, in return, gain economic value from AI.

If you answer no to any of these questions, then perhaps a simpler, less costly solution would suffice.


  • Does the business problem directly impact our company strategy?

An investment in AI can change the ways of working and require some upfront cost of time and resources, so invest in AI when the problem is inhibiting your business from achieving desired results.

  • Can we see value from our investment in 1 month?

Working agile to find early wins is critical when making a large investment in AI to ensure that the team is on the right track to creating something useful.

Problem state:

  • Do you have ample organized and structured data about the problem for AI to analyze?

AI works best when there’s a lot of structured and organized data (e.g. think all the data that financial markets produce everyday); that said, AI can also find patterns in your data to find a structure that human eyes may not perceive.

  • Do simple rules and formulas struggle to work effectively due to the difficulty of the problem?

Sometimes simple math formulas and heuristics work perfectly fine in providing you with the answer you need; Using AI when statistics will suffice is like using an industrial kitchen to make one peanut butter & jelly sandwich — it’s overkill.


  • Will the solution still be relevant in 3–5 years?

Deploying an AI solution can be time consuming and costly, so the investment should still payback in years to come.

  • Will a majority of our employees and / or customers benefit from the solution?

Any large investment should create a large return on investment, so invest in AI whenever it can impact a majority of your workforce and / or customers.


  • Can we afford to support and maintain the solution?

AI applies to problems that are inherently messy, meaning a high risk and high maintenance cost.

  • Do we have the right resources to tackle this problem?

Finding relevant AI engineers can be challenging, and working with 3rd party AI experts can be costly, so it’s good to ensure that the right resources are in place before investing in an AI solution.

Image for post
Source: Sana Labs

If you’ve answered yes to all the questions, then an investment in AI can make sense. If you’ve answered no to any of the questions, then consider exploring a simpler approach using a rules-based engine or a heuristics-based approach. The reason for this critical upfront evaluation is that AI applies to problems that are inherently messy. Therefore, it makes sense to choose problems wisely and to strive for early wins.

Once you’ve passed the first step, then the second step is to reason around whether to build your own solution or to buy a solution from a 3rd party.

  • If the solution will significantly enhance your core business, it can give you significant competitive advantage, and you currently have or can easily acquire the talent in-house, then build.
  • If the solution impacts supporting parts of your business, and you currently don’t have or cannot easily acquire the talent in-house, then buy.

The benefit of buying over building is that you can make significantly less investment up front and you can focus on more important aspects of your business.

Image for post
Source: Sana Labs

Using AI in your learning agenda to create economic value

Once you’ve identified an outcome (for simplicity’s sake, let’s refer to the “outcome” as an output) that you can solve by investing in AI, your team has to identify the inputs for AI to generate the desired output. Currently, AI creates economic value in one type of way: an input is augmented, mapped, or analyzed, and an output is created.

The economic value created from AI depends on the improvement that AI brings to the input-output transformation. The improvements are through automation or augmentation of workflows, which translate to a reduction in time, a reduction in errors, and / or an improvement in productivity. The output (or outcome) could be a better onboarding experience, enhanced compliance training, or a more relevant product training.

Image for post
Source: Sana Labs

Let’s visit a common challenge that some CLO’s face when driving their learning agenda, as an example of how AI used in the learning space can create economic value.

Challenge: Creating up-to-date relevant content takes significant time and investment from my team. The industry standard currently suggests that 80–120 hours is spent creating a 1 hour e-learning course.

  • Desired output: An automated, intelligent course authoring workflow for creating and updating relevant content in 50% less time.
  • How AI delivers value: Recommend images, automate content design, and generate assessment questions to solve a majority of the content creator’s heavy-lifting efforts.
  • Inputs: Data that captures attributes of content and images, and the author’s historical behaviors and preferences.

Here’s another example of how AI used in the learning space can create value for your organization.

Challenge: Learners in our organization complain that our learning content is not engaging, too easy or too difficult, and become unmotivated and disengaged. They ultimately leave our company to find a new job that invests more in their professional growth.

  • Desired output: Deliver learning experiences that motivate learners to complete courses by ensuring that the content presented to a learner is personalized to their proficiency. Motivated and engaged learners will stay with our company, reducing severance and hiring costs.
  • How AI delivers value: Model the probability that a learner will churn (e.g. become disengaged) from the course and use the probability score to drive content personalization: providing learners with the right content difficulty for their proficiency level.
  • Inputs: Historical interaction data about the learner’s proficiency of the subject matters and the difficulty of the content.

Understand the problem and choose the solution accordingly

In summary, CLO’s should understand the problem space using the decision criteria before investing in AI to solve challenges faced while driving their learning agenda. Once a suitable problem has been identified, then the CLO, together with experts in AI, can either build or buy an AI solution that can deliver real economic value.

In one example of AI driving business value, Sana Labs helped nurses learn critical care education 37% faster using Sana compared to traditional teaching methods (e.g. using an LMS, reading PowerPoints and PDFs, watching videos). 37% faster learning translated to an annual opportunity benefit of $1.4m for the hospital in addition to giving back 6 hours to each nurse that would be used improving the well-being and lives of their patients.

Partners like Sana Labs can help provide AI technologies such as those found in Sana, the personalized learning platform, to help CLO’s realize value from AI and improve learning outcomes. To learn more about AI in the business context and to understand more details of how and when AI can be applied effectively, try out The essential AI handbook for leaders on Sana, produced jointly by AI Sweden, Peltarion, and Sana Labs.

Finding relevant AI engineers can be challenging, and working with 3rd party AI experts can be costly, so it’s good to ensure that the right resources are in place before investing in an AI solution

Spread the word

This post was originally published by Jon Lexa at Medium [AI]

Related posts