Most people hear “AI” and think of tools that spit out text, images, or code. That is useful, but it is not the same as helping people learn. At Drillster we focus on the kind of AI that supports real learning, the stuff that helps people remember, apply what they know, and perform better on the job.
Below are the practical ideas from the webinar: how memory works, why short practice sessions and fast feedback matter, and how we use adaptive learning plus constrained generative AI to create training that actually sticks.
The problem: we forget fast
Learning starts with attention. If people are not paying attention, nothing sinks in. Even when someone does pay attention, information first goes into short-term memory, and if it is not used, it fades quickly.
Research shows people can lose up to 80 percent of what they read or see within a week if they never use it. That is the forgetting curve. For critical topics such as safety or security, that is a real risk: knowledge fades, performance drops, and mistakes happen.
So the question is simple: how do you get important knowledge and skills into long-term memory so employees can pull them out instantly when needed?
Three people, three levels of mastery and one goal
Imagine three customer service managers:
- Sarah: ten years at the company, always refreshing her knowledge, lots of long-term anchor points.
- Bob: a few years of experience and some training, decent but not as deep as Sarah.
- Tom: brand new, still onboarding.
The aim is for all of them to recall and apply the same core skills when they face a real situation: a difficult customer, an objection, a process exception. The trick is to train the brain so it knows exactly where to find the right response.
How learning actually happens
Passive activities such as reading, watching, and listening are weak for retention. Active retrieval is where learning becomes durable. That means:
- answering questions,
- solving dilemmas, and
- reacting to realistic scenarios.
Every time someone tries to recall an answer, the brain strengthens that memory. Getting feedback right away matters too. The instant correction window is when the brain is most ready to absorb the right information. That is the core idea behind Drillster’s approach.
The Drillster approach: short, active, personal
We combine neuroscience with AI and practical learning design. Here is how it works in practice.
1. Microlearning and drills
Content is broken into small, focused chunks learners can practice in short sessions, for example 5 to 10 minutes on the commute or between meetings. Each chunk is followed by questions or scenarios we call drills. This turns passive exposure into active retrieval.
2. Immediate feedback
When a learner answers incorrectly, they get an immediate correction and explanation. That moment, when the brain realizes it was wrong, is the most powerful learning opportunity.
3. Adaptive repetition and smart spacing
Drillster tracks each person’s learning history and uses didactic AI to decide which items need repeating, when, and in what form. New learners get more frequent refreshers, experienced people get less frequent ones. Over time, the gap between repetitions grows, which strengthens long-term retention.
4. Nudges before forgetting
Instead of waiting for skills to drop, the app predicts when a topic will fade and nudges the learner to brush up just in time. That flattens the peaks and valleys and turns one-off training into steady, year-round readiness.
What that looks like in the product
From the learner’s view you will see a topic such as cybersecurity. The interface shows current proficiency and a forecast of what will happen if you stop practicing. Each learning item has microlearning content followed by drill questions. Correct answers raise proficiency, incorrect ones trigger targeted repeat practice. Over time, learners efficiently approach full proficiency.
Managers see aggregated readiness and can target interventions where they matter most.
Creating good content quickly: Question Crafter
Content is only useful if it focuses on what people really need to know. That is why we built Drillster Question Crafter, a way to generate adaptive learning content fast using the documents you already trust, such as policies and playbooks.
Here is the flow:
- Upload an internal document, for example leadership guidelines.
- Tell the tool who the audience is and what cognitive level you want, based on Bloom’s taxonomy: remember, understand, or apply.
- The system analyzes the document and extracts the need-to-know learning elements. You can edit those elements.
- Choose how many equivalent questions per learning element so repeat practice uses fresh variants rather than the same wording.
- Choose distractors and language level, useful for multilingual audiences.
- The tool generates question sets that follow didactic design rules and checks that it only used the uploaded document as the source of truth.
The result is ready-to-use adaptive drills that map precisely to your internal guidance and not to the whole internet.
Little but important design choices we built in
- Focus on what to do rather than what not to do, because learners must recall correct actions in the moment.
- Use equivalent questions instead of exact repeats, to avoid recognition effects.
- Offer language level control so content is accessible to the intended audience.
- Run a second-look validation to ensure generated content sticks to your documents and didactic principles.
Generating a set of drills from a typical document takes from a few dozen seconds to a couple of minutes. After that you can review, tweak, tag, add media, and release.
Why this matters to the business
When knowledge and skills are anchored, employees make fewer mistakes, stay safer, work more efficiently, and perform better. That shows up in fewer incidents, higher productivity, and ultimately better business results.
Because content can be generated quickly from your own documents and practice paths are personalized, you do not need massive, expensive rollouts to see value. A focused pilot for one role or one process can already change outcomes.
Quick checklist for getting started
- Pick one critical role or topic such as safety, customer responses, or cybersecurity.
- Gather the key guidance documents you want to lock into long-term memory.
- Use microlearning plus drills to create short practice sessions.
- Enable immediate feedback and adaptive spacing.
- Run a pilot with a small group, measure proficiency gains and behavior changes, then scale.
Final thought
Generative AI can create content fast. Didactic AI makes that content work as learning. Put them together, but anchor the content to your rules and your reality, and you will end up with training people actually remember and can use when it counts.
If you want a walkthrough of Question Crafter or a short pilot plan for one job family, we can sketch one out together.
