Last month, StudyABAExam.com opened its doors and welcomed exam candidates preparing for what very likely could be the most challenging exam of their lives. We have been operating for about a month now, so it was a good time to put my ambitions for StudyABAExam.com in writing.

But first, if we haven’t had the pleasure of meeting yet, Hi 👋, I am Jake.

As a professional, I have one core principle that has repeatedly influenced my decisions: I want to help others. It's why I got into behavior analysis in the first place. I was extremely fortunate to be exposed to high-quality service delivery early on, which showed me the degree to which behavior analysis can create a meaningful impact on those we serve. That same core principle (i.e., helping others) is what led me to learn data science and, in turn, software engineering. I wanted to help more people. I wanted to scale my impact to hundreds and thousands of individuals.

There is a saying I really like in software engineering: "write once, run everywhere." I revisit that saying a lot, and here is how I think about it. If you are a house builder and you build a house for someone, that is incredible. You brought something useful into the world, something that didn't exist but was needed. Now imagine that instead of building a house one at a time, you could build it once and have it serve everyone who needed one, all at once. That is the idea behind "write once, run everywhere." Data science works the same way. Instead of a home, you are building a model, but the principle holds true. Build it once, and it can help not just one person, but hundreds or thousands at the same time. The scale of impact grows, but the mission stays the same.

But enough about me and my journey... What about yours?

Well, if you look at the landscape of test preparation and the chatter from individuals who are undertaking this journey, you will likely see stories about multiple test takers who have used different materials to prepare. They post reports of jumping around from one platform to the next, investing hundreds, if not thousands, of dollars into exam prep. They make attempts to try and contextualize the information they have (i.e., “I scored this on this exam, am I ready?” or “Here are my scores for X, Y, and Z. What should I do?”). These sentiments are often fraught with frustration for those who have spent many years altering their behavioral patterns to meet eligibility criteria to even sit for the exam. But, for one reason or another, they aren’t able to meet the criteria for passing their exam.

Now, I don’t know about you, but I am a behavior analyst. I see that, and I see the reflection of the environment. Not the failure of the individual.

These sentiments, and many others, are why I believe StudyABAExam.com offers such an amazing opportunity. And how AI can help solve the problem of understanding when you are ready.

Before I get into why AI is so well-positioned in this space, I want to make one thing crystal clear. When people talk about AI these days, they are typically referring to large language models. Labs such as OpenAI, Anthropic, and Google have brought AI to the masses, which is great. The models they provide represent a specific flavor of AI, namely, large language models. In fact, I had generative AI as a feature in StudyABAExam.com at launch, and it was a pretty awesome system, and personally, one of my favorite things I have built. But I took it out. Killed it. Why? Because people preparing for the BCBA exam don't need stochastic systems, and they don't need generative questions. Instead, they need prescriptive preparation. They need a signal.

Lucky for us, AI is more than generative. There is a rich history of AI that is independent of the current mainstream models. The space I am referring to is supervised machine learning. The methods and opportunities that exist in this problem space, to me, reside within those offerings from AI. There are statistical insights that we can draw from users of our platform that can give a signal to others.

So what does this actually look like in practice?

Well, to start, we started by aligning our content with the test content outline. Users on StudyABAExam.com are getting real value from material that is tightly aligned with the 6th edition test content outline. That alignment matters. The exam has a blueprint, and preparation should reflect that blueprint.

Beyond content alignment, we made a deliberate decision to cast a wide net. By offering mock exams at a fraction of the price of competitors, we accomplish two things at once. First, we lower the barrier to entry for candidates who have already spent significant money on their journey to certification. Second, and this is the future I am most passionate about, we are able to analyze a wide swath of questions from a statistical standpoint. That means running item analyses, examining how distractors perform, and understanding which questions are doing their job and which are not. A question that everyone gets right is not a useful question. A question that everyone gets wrong might be a poorly worded one. Either way, we are going to let data be our guide.

Advancing past statistics into prediction means we need to have control over every layer of the system that produces our data. That is why building this platform ourselves was never just a product decision; it was a data decision. We get to decide exactly the level of aggregation at which we collect information. Those decisions are feature engineering decisions. Every architectural choice we make today is a choice about what models we will have access to tomorrow. We constructed the pipeline with this end state in mind, so that when we have the volume of data we need, we are not scrambling to retrofit a machine learning system into a structure that was never built to support one. The foundation is already there.

But if I am being completely honest with you, we are still early. Right now, we are sitting in a data waiting room. The most precious resource we have is not the technology, it's not the content. It is the data we get back from people using the platform. Every exam attempt, every question answered, every pattern of strength and struggle is a signal. And the more of those signals we collect, the sharper our ability to tell a candidate exactly where they stand and exactly what they need to do next. That is what we are building toward. And we are just getting started.

If you are an individual preparing for the exam or a stakeholder who is in a position to point people in the direction of exam prep material, reach out. I would love to show you what we offer, and I am happy to meet in person and discuss further.

How AI Can Save BCBA® Exam Prep — And Why It's Not the AI You Think It Is