“The experience is designed and it is measurable.”


The intend layer (design foundation) for the ILXD model is organized by three primary topics; Learning Experience Design (LXD), Artificial Intelligence for Education (AIEd), and Analytics. In our experience, this foundation layer is where the greatest flux has been occurring. Our LXD process has been under nearly constant revision in response to advances and insights gained regarding the layers above it in the scaffold.  AIEd has evolved from a simple text-based chatbot to a full-blown intelligent enterprise. Analytics have advanced from being reflective to being a formative input for precision learning. Despite the constant changes, these three key elements of Intention have fundamental guiding principles we use in our design.

Learning Experience Design (LXD)

Introduction to Learning Experience Design

Learning Experience Design (LXD) enables education creators to better meet the needs of the student. The fundamentals of instructional design are present, but LXD improves the delivery, structure, and, most importantly, focus of the learning experience.

The fundamental shift to LXD signals a focus on the learner. Instructional design focuses the best product that can be used for instruction, but LXD focuses on the best experience the learner can have while reaching their learning goals. Behind that change is a shift to empathy, from what the learner is going to know to how the learner is going to feel throughout their learning experience.

Traditional courses are designed in modules and delivered in event-based or rigidly delivered episodic increments. LXD delivers learning experiences through content curation in chunked, a la carte pieces using innovative platforms that meet the learner where they are and take them where they need to be.

The learning methods in LXD are varied and appropriate for the outcome. Each learning element has context, so there is no fluff or ineffective learning events that don’t help the learner achieve the learning goals. A key element of LXD is learner control. The learner controls pacing, sequence, review opportunities, modality, practice opportunities, assessment methods, and reflection or processing time, among others.

Through experiential learning and a constructivist approach, we design human-centered, holistic learning experiences so that learners are empowered to create enjoyable, engaging, relevant, informative, enriching, immersive learning experiences for themselves.

The Learning Experience Designer is an instructional architect who looks at how systems operate and how each learning component fits into the overall learning experience. It starts with big picture thinking, a careful consideration of how the learner will experience the instruction we design. The designer focuses on what the learner wants to know and what they should be able to do after completion. Both the content and the user experience are designed around the learner.

LXD in Game Development

The gaming environment is fertile ground for LXD. Humans respond to experiences and learn from them, and there is no better modality for personalized learning than Game-Based Learning (GBL). It is inherently learner-centric, has the ability to reach diverse learners, and can expose learners to environments and situations they may not have access or opportunity to otherwise.

Thoughtful design can made GBL more effective to enable deeper learning and more engaged, empowered learners. The combination of effective LXD in GBL is a product differentiator that businesses would be wise to capitalize on.

For us, LXD in GBL means that we create effective, engaging learning experiences by taking advantage of the virtual reality environment. This is the “I” in ILXD. The immediate immersion and entry into flow state activates empathy and creates a different kind of learner engagement. Games can use artificial intelligence to add to the highly personalized environment, providing expertise, guidance, remediation, and knowledge extension based on the learner’s zone of proximal development.

ILXD Creation Process

As with any good design, we start with the learner first. We have to understand our audience and their unique needs before we design our experience. We serve a very broad, diverse audience, and one solution does not apply to any one set of learners.

We start by asking key questions that guide our design. We want to know about their knowledge, skills, confidence, motivation, resources and tools, and learning preferences.

  1. What drives our learners? What are their motivations and how can we access that? Why do they care?
  2. How can we connect to our learners on a personal level?
  3. How can we capitalize on the learner’s previous experience?
  4. What do they want to see in how they experience the learning?
  5. What prevents our learners from engaging with the information? What would they change?
  6. How can we connect our learners to the big ideas so they are able to construct experiences that bring the learning outcomes within their zone of proximal development?

The ADDIE development method in a waterfall of prescribed steps is too rigid for game development. Instead, we design the experience before the product based on audience needs. Design thinking guides our process through research, experimentation, ideation, conceptualization, prototyping, iteration, and testing. These steps occur in various orders and overlap. The learner co-creates the experience and we build their feedback into the product.

The reason for this is the shift in perspective from instructional design to learning experience design. In instructional design, we assume that designers and subject matter experts know best. We rarely audience test or involve the learner in design or production. As learning experience designers, however, we start with the idea that we don’t know what the learner wants or needs and work through our design process, including compulsive play testing, with the goal of learning the most effective design structure will be.

Application of ILXD in GBL

We have created several games and simulations that exemplify the efficacy of ILXD in GBL. In Cold Case, the learner solves a crime through interrogations, evidence collection, and lab tests and is guided by a Sergeant to reflect, review, and complete processes. The system empowers the course instructor to facilitate learning through the environment. In Take Control, the learner, guided by an AI agent, determines their own conflict management style and uses that style to manage conflicts in an office environment. Their own choices determine their experience, what challenges they encounter, and how well they meet the learning objectives. Our Microscope simulation provides learners with unlimited opportunities to learn and practice microscopy fundamentals. AI supports the learning and instructors can use the environment to facilitate learning outcomes. Our Pathology Engaged Learning Objects (ELOs) immerse learners in a hospital room and enable them to experience each body system through chart analysis, medical tests, and exploration. The content varies for each ELO, and the spaced repetition of key concepts ensures retention and reinforces application of knowledge.

AI has been an effective element of the learning experiences we have designed. It is a fundamental element in designing for the ILXD Pyramid, and understanding its capabilities and uses is key to successful ILXD creation.



Artificial Intelligence for Education (AIEd)

Introduction to Artificial Intelligence

AI involves computer software that has been programmed to interact with the world in ways normally requiring human intelligence. This means that AI depends both on knowledge about the world and algorithms to intelligently process that knowledge (Luckin, Holmes, Griffiths, Forcier, 2018).

In AIEd, this knowledge about the world is represented in three key models: the pedagogical model, the domain model, and the learner model.

An AIEd system that is designed to provide appropriate individualized feedback to a learner requires that the AIEd system knows something about:


• Effective approaches to teaching (which is represented in a pedagogical model)

• The subject being learned (represented in the domain model)

• The learner (represented in the learner model)

AIEd functionality is often described as Machine Learning and Deep Learning, which can be understood as subsets of Artificial Intelligence. This relationship can best be illustrated with a diagram of concentric rings, where Machine Learning is subset of AI, and Deep Learning is a subset of Machine Learning.

Machine Learning is the practice of using algorithms to sort data, learn from it, and make a determination or prediction about something in the world.

Deep Learning is typically a practical application of Machine Learning. Driverless cars, medical diagnosis, and shopping recommendations are all examples of Deep Learning.

What AIEd Can Now Deliver

Affordable, accessible AIEd technologies and tools are now available to support 1:1 learning at scale. Every learner, in every subject, can learn what they need to know, at their own pace, in a context that is most meaningful to them, working in collaboration with classmates that are ideally matched to suit their needs.

AIEd can provide each learner with a personalized intelligent tutor that accompanies them throughout their entire program of study, accessing real-time analytics that measure learning outcomes to shape each learning experience real-time to best suit the learner.

AIEd can provide each learner with personalized experiences that precisely address their individual needs and progress. It can provide learners with specific, actionable, relevant feedback in real-time for every challenge and task along the way.

AIEd is capable today of providing inspiration, communicating with emotion, listening to the learner’s voice, reading their facial expressions, and responding to the learner’s successes with a continuum of ideally suited levels of challenge.

Our initial AIEd application was a text-based ChatBot in an online game for a university English course. Our first playtest immediately revealed two important findings regarding this feature:

  1. Gameplay soon came to a complete halt for the entire class. They became engrossed in conversation with the ChatBot. This deep engagement did not dissipate, they continued dialogue with the ChatBot until we finally had to intercede in order to complete the playtest.
  2. With few exceptions, the students were asking the ChatBot probing questions.  They were most captivated by exploring the knowledge boundaries of the AI, to see if they could ask questions that it could not answer and they were sharing with their neighbors what they were observing.


A few months later, a new AIEd that we deployed was a non-player character in a VR soft-skills training product for professional education. This NPC, named “Victoria”, spoke to the learner in response to dialogue choices that they made. Victoria had been designed to misbehave in a scenario-based learning challenge. Our first playtest with this AIEd revealed that learners left the scenario with a strong emotional response to the NPC, talking about Victoria as if it were human. We later demonstrated the product at a large education conference in our expo booth, and by the second day of the conference we had attendees coming to the booth because someone had told them about “Victoria” and they wanted to try the VR experience and meet her.

Scenario-based experiences using dialogue with AIEd non-player characters is a primary focus for our products. Our small in-house team designs, develops, and delivers these intelligent immersive experiences using off-the-shelf tools from IBM, Google, and the Unity game engine. An important aspect that our learning experience designer has been focusing on for these experiences is imbuing a wide range of emotional responses in the dialogue designs that comprise the ‘brain’ that powers the AI characters. In this way, we constrain the possible responses for the AI characters to only the content and emotion desired for the scenario based upon the decision-making of the learner.

One of our newest intelligent immersive learning products is an enterprise effort using Scriyb, an AIEd product resulting from research conducted by the Virginia Serious Games Institute. This AIEd tool optimizes and supports collaborative learning groups.  Each learner has an intelligent tutor that gains insight about them and assigns them to a group most suited for their needs. Scriyb adjusts these assignments dynamically as the term progresses, to maintain the optimum group for each learner.

An argument against the use of AIEd that we occasionally hear is that it will replace professors. Our intention is the inverse, we see AIEd empowering professors. In large courses and programs, one size can’t fit all. Inevitably, some students fall behind and eventually fail. An individual professor can’t possibly engage in a the 1:1 interaction that is most effective. This is especially true for online courses. As a result, the professor has little choice but to use the lecture-style approach for the large group - which is the least effective way to learn. Unfortunately, the demands of assessing large groups of learners has also led to ‘auto-grading’, multiple choice exams that are extremely poor instruments for measuring learning outcomes beyond the basic recall of facts.

In providing 1:1 personalized, real-time learning support for each learner, and by providing real-time learning assessments, AIEd empowers the professor to assist learners when they need it, precisely targeting specific areas where they need that assistance the most. The AIEd empowers the professor to use a wide range of assessments based on what the learner finds most effective. Also, the AIEd assists the professor in collaborative group assignment and facilitation, again making the professor available to assist when and where it is needed the most.

Building and maintaining trust in AIEd is an important new dynamic that is critical in gaining acceptance for and continuing to progress and develop AIEd.

The level of trust a person has in someone or something can determine that person’s behavior. Trust is a primary reason for acceptance. Trust is crucial in all kinds of relationships, such as human-social interactions, seller-buyer relationships, and relationships among members of a virtual team. Trust can also define the way people interact with technology.

Trust is viewed as: (1) a set of specific beliefs dealing with benevolence, competence, integrity, and predictability (trusting beliefs); (2) the willingness of one party to depend on another in a risky situation (trusting intention); or (3) the combination of these elements.

Trust building is a dynamic process involving movement from initial trust to continuous trust development. Continuous trust in a learning experience will depend on the performance and purpose of the artificial intelligence. AI applications that are easy to use and reliable — and can collaborate and interface well with humans, have social ability, facilitate bonding with humans, provide good security and privacy protection, and explain the rationale behind conclusions or actions — will facilitate continuous trust development (Siau and Wang, 2018).

To build and maintain trust for AIEd in our products, we have adopted design criteria and Quality Assurance (QA) criteria for AIEd. We communicate these to client institutions and solicit their assistance and feedback for their application.

In our ILXD process and throughout production and deployment, we are using the following six metrics for regarding AIEd:

  1. Privacy and security: AIEd must comply with privacy laws that regulate data collection, use and storage, and ensure that personal information is used in accordance with privacy standards and protected from misuse or theft.
  2. Transparency: We must provide contextual information about how AIEd systems operate so that people understand how decisions are made and can more easily identify potential bias, errors, or unintended outcomes.
  3. Fairness: When AIEd systems make decisions or recommendations they must be fair. We must understand how bias can affect AI systems.
  4. Reliability: AIEd systems must be designed to operate within clear parameters and undergo rigorous testing to ensure that they respond to unanticipated situations and do not evolve in ways that are inconsistent with original expectations. Humans should play a critical role in making decisions about how and when AIEd systems are deployed.
  5. Inclusiveness: AIEd solutions must address a broad range of human needs and experiences through inclusive design practices that anticipate potential barriers in products or environments that can unintentionally exclude people.
  6. Accountability: When we design and deploy AIEd systems we must be accountable for how the systems operate.

Part of building trust includes responsible treatment of the learner data that AI provides. Our standards ensure we can be this data conscientiously to enhance the learning experience.


A common definition of Learning Analytics is the measurement, collection, analysis, and reporting of data about learners and their contexts for the purposes of understanding and optimizing learning and the environments in which it occurs.

Learning analytics uses the data associated with a learner’s interactions with content, other learners, and the educational institution to make decisions and evaluations about teaching practices, personalized content, and needed interventions for learner success. The field draws on and integrates research and methodology related to data mining, social network analysis, data visualization, machine learning, learning sciences, psychology, semantics, artificial intelligence, e-learning, and educational theory and practice. Learning analytics focuses on the interpretation of the educational data from a learner and teacher orientation. This places as much emphasis on understanding the pedagogical context from where the data is derived as developing statistically robust interpretive and predictive models.

Analytics Cycle

Learning Analytics are developed in a multi-step, cyclical process of data collection and pre-processing, analytics and action, and post-processing. Data collection and pre-processing involves the gathering of educational data from different learning tools or applications and preparing and translating it into an appropriate format.

The analytics and action phase denotes the actual application of analytic methods (e.g. structure discovery, relationship mining etc.) to extract meaningful patterns and information from the data and to make use of the obtained results, (e.g. visualization, feedback, recommendations, adaptation).

Post-processing refers to the idea of continually improving analytics, by refining analytics methods, using new methods, including new data sources, etc..

In the context of an intelligent immersive learning experience, learning analytics empowers AIEd through the use of intelligent data, learner-produced data, and analysis models to discover information and connections, and to predict and advise on learning. That data mining and analysis enables AIEd to provide real-time assessment of learning experiences. Predictions can be drawn on the basis of patterns of behavior as each learner follows in their individual learning pathway, and it can find hidden patterns in interactions that can be meaningfully interpreted and then fed back to
learners in a way that supports their learning.

The intend layer of the ILXD pyramid is essentially the fundamental design phase. The elements are designer/developer-focused and structural. The use and process of learning experience design creates the learner-focused experience. Employing powerful tools such as AIEd and analytics flesh out the experience to make it truly engaging and immersive.