Using context-consideration framework for EdTech AI products
Identifying AI ideas with high market-fit chances
What it takes to build AI products that have a high chance to reaching market fit? That’s what we will be exploring in this writeup — but specific to EdTech.
Only recently I learnt about context-consideration framework. In reference to our topic — AI models/products — ‘context’ refers to the volume of abstract concepts that the model is required to know and ‘consideration‘ refers to the amount of effort our user/customer takes to make a decision for the use case.
Let’s take an example from the EdTech arena — remember the time when you completed 12th grade and started looking for the perfect college for yourself. How important was this decision for you and your parents? One of the most important life decisions right? Why so? Because you were going to only take it once, or because this decision was going to define the next few years of your life, or it was very difficult to reverse once taken or irreversible in some cases. So deciding a college was ‘high consideration‘ for us.
How did you make this decision? Since it was high consideration, you probably searched for an expert to help you with it. Expert of what? Expert in graduate education — this expert could have been an education councillor, or someone with experience of college hunting, or it could have been raw college placement/results data. All these expert solutions have context specifically related to graduate education — in depth understanding of limited concepts — hence ‘low context’.
If we had to build an AI product for this use case, it would have been a model trained on graduate education data, catering to niche highly considerate user persona(s), well priced since the user spends huge effort (since money in exchange of low effort is a proven business model). Someone who has access to ‘graduate education data‘ is better placed to test and succeed in implementing this idea.
The above paragraph derived from our context-consideration understanding, answers many of our questions. These answers will tell us what are the chances of being market-fit if built by us/our team.
Targeting/GTM: Should be niche, just one or two personas, (why niche? because user has high consideration)
students aged 17-18 years
parents aged 40-52 years with kids
Pricing: Moderate to High price based on segment income and competing product prices (manual products/web products).
Moat: High tech feasibility if we/our team has access to ‘graduate education data‘. ChatGPT like model with extremely high context is not much useful over here without our limited context data. Low context AI products hence still can create very strong entry moats.
Though the above thinking process gives us some points to start with, only qualitative and quantitative data from the market can prove the fitness of our product — hence launching MVPs will be core. Since we have an idea of the target, the confidence that money market exists and some degree of moat is possible — this is worth an MVP.
If you found this brainstorming worth it, consider subscribing: