Handle with care: AI
Five years ago, when we launched OMY! Sports, we were careful about how openly we talked about AI — we didn’t want to put off users who were used to working with a human coach. Today, the situation has changed, but the caution remains. Only now for a different reason: expectations around AI have become extremely high.
As is often the case, reality lies somewhere in between.
Where the expectation gap comes from
Naive use of open AI models — such as ChatGPT, DeepSeek, and others — often leads to two extremes. On the one hand, there’s a sense that “everything is now clear.” On the other, there’s disappointment when the answers turn out to be too generic or not precise enough.
How AI models actually work
Let’s take a closer look at how they operate. They are very good at:
- formulating ideas
- structuring information
- suggesting possible solutions
However, they primarily rely on generalized knowledge rather than a user’s specific situation. As a result, their answers are logical and helpful, but not always directly applicable in practice.
The main limitation of open AI
Modern AI models can already suggest solutions and even choose between alternatives. However, the key question is how well those choices fit your specific situation. Without understanding your level, context, goals, and accumulated data, any answer remains probabilistic rather than precise.
What makes AI truly useful
This is where the line is drawn between “interesting” and truly useful AI.
For recommendations to become practical, it’s not enough for a system to “know everything.” It needs to work with individual user data and apply a clear and proven decision-making logic. Importantly, data alone does not create value — value emerges only when that data is interpreted through a methodology.
Open vs Product AI — what’s the difference?
In this sense, the difference between open AI models and product-based systems becomes especially clear. Open models provide breadth — they update quickly, explain well, and cover a vast range of topics. Product systems, on the other hand, operate with specific user data and goals, allowing them to deliver more actionable recommendations.
Individually, both approaches have limitations:
- the former lack precision
- the latter lack breadth
Where real value in AI emerges
Real value comes from combining three elements:
(1) the breadth of open models + (2) the depth of user data + (3) a decision-making methodology
It is the methodology that connects data and recommendations into meaningful, actionable outcomes.
How it works in OMY! Sports
At OMY! Sports, we’ve brought together three key components:
- the capabilities of open AI models
- analysis of individual athletic data
- a decision-making methodology based on professional coaching experience
These are not separate elements, but a unified system that operates across all levels of the product:
- in conversations with Rob
- in communication with a coach
- in the training calendar
- in settings
- and at the backend level, where training plans are created and continuously adjusted
This is what allows us to turn answers into actions — not just information.
What’s next
The continued evolution of AI, including the rise of AI agents, will further improve the user experience. Interactions will become faster, simpler, and more natural. However, the core principles will remain unchanged: working with individual data, applying structured decision logic, and combining open and product-based AI.
Why it matters
AI becomes truly useful not when it “knows a lot,” but when it helps you make the right decision in a specific situation. Understanding this helps set realistic expectations, choose better products, and ultimately get real value — not just answers.
You can learn more about the OMY! Sports methodology in the AI section of our website.