AI-Powered Discovery
Transforming learner motivation & achievement with a little help from AI
The “D” of DOMS™️ - my evidence-based learning design process - stands for Discovery.
The goal of learner discovery is to deeply understand the “why” of your learners. Research shows that by doing this, we can optimise our designs for learner motivation and, as a result, learner achievement.
In this week’s blog post, I explain how Discovery can transform learner motivation and achievement and share a Discovery edu-mega-prompt to help you transition from basic learner analysis to game-changing learner Discovery.
Let’s go! 🚀
What is Discovery?
I borrowed the term Discovery from the worlds of Design Thinking and Product Development.
In both of these contexts, Discovery is all about defining about user pain and user delight. In a nutshell, Discovery asks:
Why am I building this product?
Who am I building it for?
If I build it, will they come?
Traditional instructional design models like ADDIE make some attempt at discovery in the analysis stage. However, my research shows that instructional designers’ analysis is typically limited in four ways:
It focuses on an analysis of the needs and motivations of the course creator (e.g. the organisation or professor) rather than the learner. We tend to focus most on “what I want to teach” and far less on what the learner wants to learn, how and (critically) why.
It focuses on practical information about learners. Our analyses tend to focus most on things like the time that learners have to dedicate to a learning experience without asking first why they would dedicate any time to the experience at all.
It fails to consider key qualitative insights which enable us to optimise learning designs for learner motivation & mastery of knowledge, skills & behaviours.
The result? We rarely optimise the design of learning experiences for learner motivation or, as a result, the mastery of learning objectives.
TL;DR: We work on the assumption that “if we design it, they will come and they will learn”, but without surfacing and drawing on the right sorts of learner insights at the start of the design process, we face a critical triple threat:
learners don’t show up at all;
learners show up, but don’t stay;
learners show up and stay (e.g. because they have to) but without intrinsic motivation or genuine interest, they fail to engage in a way that’s required to achieve our intended objectives.
The Learning Science Bit
The case for a shift from traditional methods of analysis to more robust Discovery is compounded by a large body of research on why learners learn.
Here’s a brief summary of some of the most important Discovery research:
Positioning is Everything: Genuine learner interest in a topic leads to intrinsic motivation, resulting in increased engagement, retention, and achievement. (Malone & Lepper, 1987)
Map to Learner Pain & Aspirations: By aligning the learning experience with learners' personal and professional beliefs, drivers, and values we enhance intrinsic motivation, engagement and achievement. (Ryan & Deci, 2000)
State the Why: When we clearly explain the immediate relevance and value of the learning experience, such as career progression, we motivate them to complete a learning experience. (Knowles, Holton & Swanson, 2012)
ZPD is Queen: When we identify and design within our learners’ Zone of Proximal Development (ZPD), we optimise the learning experience for motivation and mastery. Ensure learners experience a level of challenge that fosters engagement. (Vygotsky, 1980)
Eliminate Friction: When we minimise obstacles and design learning experiences that fit seamlessly into learners' daily routines, allowing flexible completion within their work schedule. (Bersin, 2018, Veletsianos et al., 2021)
TL;DR: By understanding who our learners are, what they want & need, we not only increase their motivation to learn but also optimise for the achievement of intended outcomes.
Shifting from Analysis to Discovery
I’ve spent the last few weeks talking to ~100 learning designers about why they don’t dedicate more time to Discovery. The answer is two fold:
Time: “Uncovering in-depth insights is time consuming. There is very limited time in my process for running surveys & interviews, then analysing all of the data.”
Expertise: “To be honest, I wouldn’t know where to start in terms of writing the right questions to generate this sort of data… I’m not trained in data analysis, so I’m not confident that I would be able to handle the information I generate.”
So, the question I have started to explore is: can we use AI to help overcome the barriers to Discovery and help learning designers to supercharge their design process?
It’s a work in progress, but what I’ve learned so far is that AI can’t do Discovery for you, but with the right prompting and due validation, it can help a lot.
Let’s look at an example.
AI-Powered Discovery
Given the limitations of their training data, non-internet-connected AI tools like ChatGPT 3 & Claude 2 don’t have access to enough up to date data about your learners to be able to generate reliable Discovery insights. You can try using these tools, but the outputs require so much validation that the benefits are pretty limited.
Internet-connected AI tools like ChatGPT 4 + BrowserOp Plugin & Bing still don’t know enough about your learners to be able to “do your Discovery for you” but, prompted in the right way, they do a pretty solid job at getting you started and accelerating your process.
The key thing here is the quality of the prompt. By combining learning science research with the efficiencies offered by AI, magic can happen….
Here’s the prompt I used:
Here’s the result:
⚠️ WARNING: As ever: AI needs you! Hallucination is real. Validate everything AI produces.
You can find a copy and paste version of the prompt here.
That’s all folks! I’d love to hear if and how this prompt works for you. Let me know by commenting below (paid subscribers) or on my related LinkedIn post.
Stay tuned for more updates on my exploration of AI-powered Discovery & design.
Happy Discovering!
Phil 👋
PS: If you want to learn more about how to embrace AI and try it for yourself, my AI-Powered Learning Science Bootcamp might be for you! The next cohort will kick off in September - you can register you interest now.
If you enjoyed this, you might want to follow me on LinkedIn and/or (hot off the press!) TikTok.