The future is unwritten, but we know with certainty that artificial intelligence (AI) will transform how we access, design, engage with and apply learning. As the promise of AI materializes into reality, it’s less clear where we are in that transformation and how best to engage it. Answers to even basic questions remain murky: What qualifies as an AI-enabled solution? How sophisticated is the algorithmic logic under the hood of existing AI learning technologies? What’s the timeline for enterprise-strength availability and adoption of emerging AI platforms?
These uncertainties are contributing to an acute “wait calculation” for many learning organizations. The term “wait calculation” was originally used to describe a dilemma associated with space travel: Is it better to invest in launching a long-range interstellar mission now or to wait for the invention of more powerful energy sources (which might allow a later launch to overtake the original probe mid-flight)?
In the less rarefied context of human learning AI, the wait calculation involves determining the right time to avoid the two extremes of indefinite paralysis (failing to commit to any technology for fear of being wasteful) or premature adoption (diving in too soon and wasting budget and resources on a low return). In the time/value trade-off at the heart of the decision, are there more benefits to being an early adopter of an existing technology or to holding off for the next wave of features and providers?
Even this framing oversimplifies the issue, because AI in learning is not a single product or technology. Human learning AI will be transformational because it will be comprehensive and, eventually, factor into every learning technology. It will include progressively more powerful, expansive and integrated platforms and calculations.
There are five dimensions of human learning AI, each of which has its own wait calculation based on its potential benefits to learning organizations and the businesses and learners they support:
AI-powered recommendation engines and adaptive algorithms match learners to the right content and playlists, connect coachees to the right coaches, enable intelligent searching, and enhance self-directed learning by making suggestions.
These capabilities already exist in numerous platforms, such as the learning management system (LMS), learning experience platform (LXP), coaching platforms and bot-based search platforms. They will continue to improve in accuracy and relevance over time, so it’s a good idea to integrate them now into your federated search platform.
To do so, work closely with business groups to determine skill needs, and map learning content against those skills. Then, provide diverse coaching resources order to optimize the platform for a range of backgrounds, preferences and needs.
2. Real-time Feedback
AI can monitor learners’ voice, face and body motions in the context of an interaction (and sometimes location) to provide in-the-moment feedback and cues about how to adjust their behavior. More advanced logic will factor in the responses and actions of another person to provide customized recommendations based on the ongoing back-and-forth flow.
Platforms have already emerged with these capabilities to support contact center representatives and other roles involved in remote customer interactions and location-specific learning. When implementing this type of platform, identify the job roles, interactions and moments of performance that real-time corrective feedback could improve.
A variant on this dimension of AI is the rise of deep analytics and reports from simulated and avatar-based practice scenarios. With these capabilities, you can seamlessly integrate role-plays and other AI-assisted learning elements into the flow of work.
3. Performance-driven Learning and Nudging
With this capability, AI defines an individual’s learning needs based on their current performance and the skills they will need in the future. The fully realized power of AI (and the one that inspires both utopian and dystopian forecasts) is that it will be able to predict people’s actions, foresee their upcoming performance issues and detect their skill needs — before they do and better than they could.
Smart bots will take the guesswork out of learning (and the biases and human shortcomings associated with self-directed learning) to guide learners toward learning and development decisions that achieve optimal outcomes. Smart learning journey recommendations that dynamically adjust over time are now available. With these platforms, you can redefine the balance between structured/guided and unstructured/self-directed learning experiences.
Existing AI tools can also connect with productivity and work platforms, such as email, calendaring and customer relationship management (CRM) platforms, to analyze learners’ entries and communications, nudging them toward improved frequency and quality of behaviors (e.g., improved feedback for team members).
Prepare for these benefits by connecting your learning strategy to performance measurement and talent management. These features will extend into AI-assisted course creation and content aggregation (constructing learning experiences on the fly, based on instantaneous data cues regarding how a learner engages) that can also be part of a performance campaign.
4. Team-based Learning
AI can organize the right participants into learning teams. More powerfully, it will take us beyond personalization to team-based capability building, where individual learners develop skills and mindsets in relationship to colleagues on their project and functional teams. In the future, this capability will include integration with workflow tools and dynamic adjustments based on the relative activity of team members (e.g., if one is not staying up to date on technical advances) and changes in team composition.
While still nascent, there are group-based learning platforms with recommendation and sharing capabilities and insight dashboards for team leaders. This feature is one to watch for as organizations realign around dynamic cells, internal influencers and cross-functional tribes. Start planning how to dynamically correlate learning with staffing to define a strategy for optimizing people resourcing, and align with business groups on team roles and capability intersections.
5. Deep Analytics
AI captures and analyzes data to determine learning impact and effectiveness across experiences. It provides insights into performance and capability gaps, which learning content or programs works best (and why), and recommendations for how learning functions need to prioritize and evolve. The impact of human learning AI will be felt not just at the individual and team level but also at the organizational level (and potentially across industries and learner populations through aggregated data lakes).
Refine your current design and development processes to incorporate analysis, and reprioritize expert involvement and resource allocation based on areas of demand and low proficiency. This dimension of AI is the furthest out in the wait calculation, but learning functions should start preparing now by implementing clean data input protocols, internal analytical capabilities and consistent measurement standards.
There’s one more factor to consider regarding wait calculation: Technology enhancements and innovation do not occur at an even rate over time. They are better illustrated as a flattening staircase — rather than a straight line or a simple curve — with irregular moments of breakthrough and rapid value addition. Consequently, there are periodic points when innovation slows down and the advantages of waiting decrease.
The five dimensions of AI described here are at different stages of progress, with matching furthest along, followed by real-time feedback, performance-driven learning and nudging, team-based learning, and deep analytics. Keep scanning the human learning AI space for upcoming feature releases and breakthroughs in adjacent functional areas, and determine whether the next wave of adoptable innovation is close or further out on the horizon.