AI WITHOUT CONTEXT FAILS - DOMAIN MODELS BEAT AI

Leverage quality data and domain knowledge for robust and context-aware AI integration

Dec 16, 2025 8m 1s

Anna Barnacka

Expert Insights

Anna Barnacka highlights the significant role of domain-specific knowledge and carefully chosen data in the utilization of AI for software development. She emphasizes 'context' as crucial, indicating that sound models take into account the domain they're operating in. Her perspective is informed by her experience in both astrophysics and Med Tech, emphasizing the quality of data over quantity and the practical usefulness of AI.

Anna makes a compelling case for:

Prioritizing quality over quantity in terms of data – a well-labeled, meticulously curated dataset is more valuable than a massive array of questionable reliability.
Harnessing domain expertise when developing AI models – for her field, deep knowledge of human physiology proves more crucial than simply deploying language models.
Integrating AI into the bigger picture – AI is not plug-and-play; it should fit alongside and work synchronously with the other parts of the system.
Employing a 'context' approach - simply applying AI doesn't guarantee success, especially in field-specific scenarios. Utilizing methods applicable to the respective field yields better results.
Building a strong, reliable infrastructure to support the AI system - maintaining synchronization and precision in data recording is essential for efficient output.

You have to provide everything around it that is needed to incorporate the technology. That's why, uh, when all partners come, we are plug and play with that whole system.

Monterail Team Analysis

Here's how to leverage context and domain knowledge in AI-assisted software development:

Prioritize Quality over Quantity: When dealing with data, it's often more beneficial to have a smaller, well-labeled and characterized dataset. Invest in curating quality data with domain relevance. It may seem counterintuitive, but it can yield actionable and trustworthy results.
Harness Domain Expertise: Understand the specific context of the area where you're looking to apply AI. The 'one-size-fits-all' approach doesn't work well for AI in field-specific scenarios. Work with domain experts or invest in training to build expertise within the team.
Holistic Integration of AI: Consider AI as a piece in a larger system, not just an isolated entity. It should complement and work synchronously with all other parts of your booking system. Plan and design your overall architecture with AI integration in mind from the start.
Pay Attention to Infrastructure: Ensure a reliable and synchronized infrastructure that can accurately record and process data is in place. This can significantly impact the performance of your AI system. If synchronization or precision is off, results may be skewed or inaccurate.
Apply Cross-Disciplinary Approaches: Learning and applying methods from adjacent or even seemingly unrelated fields can prove beneficial. For instance, techniques from astrophysics found application in Med Tech. Such cross-pollination can lead to innovative problem-solving strategies.
Adopt a Context Approach: Incorporate deep understanding of the field of application into your AI model. This can greatly enhance its relevance and effectiveness, and mitigate the risk of blind spots in AI performance.
Understand AI as an Ecosystem: Realize that AI exists as an ecosystem and not just an individual technology. This includes hardware, cloud infrastructure, data banks, and more. Designing this ecosystem is as crucial as designing the AI itself.