Many companies are currently experimenting with generic AI models, but in practice they often encounter numerous challenges. For example, an AI that has not been specifically trained with medical data may have difficulties correctly analyzing X-ray images. In this sense, Nutanix considers that the customization of AI models and their adaptation to the needs of each sector or industry will mark the development of this technology throughout 2026.
“Generative AI (GenAI) is rapidly gaining ground and is already widely used, from customer service applications to advanced data analytics. However, in practice it is proving that a one-size-fits-all model is not enough. Organizations run into limitations when AI models are not tailored to their specific sector or industry.
In this sense, Gartner predicts that by 2027 more than 50% of AI models will be sector-specific. This personalization of AI models will allow us to obtain more precise and relevant results, with solutions that are trained with data and dynamics specific to each industry,” says Jorge Vázquez, general director of Nutanix Spain and Portugal.
Current AI models fall short
At the same time, in the financial sector, a generalist model is not as effective in detecting fraud because it does not recognize the complex patterns that characterize this industry. The sectors in which having personalized AI models is a priority are the following:
- Health. AI plays an increasingly important role in the recognition of medical images, MRIs and X-rays. Personalized AI models can detect anomalies that are difficult for even doctors to identify, increasing the accuracy of diagnoses and potentially saving lives.
- Research and education. Universities and research centers use AI models for complex data analysis. Depending on the field of study, custom AI models can analyze genetic sets, simulate climate change, or study linguistic patterns.
- Financial sector. Banks and insurers rely on AI models for fraud detection and risk analysis. Algorithms specifically trained on transaction data can identify suspicious patterns that would otherwise go unnoticed, thus contributing to a safer financial ecosystem.
- Industry. In manufacturing, AI models are used for quality control and predictive maintenance. Specific AI models can detect anomalies on production lines or predict when a machine needs maintenance, increasing efficiency and minimizing downtime.
Customizing AI models will allow organizations to obtain more accurate and relevant results
Challenges in implementing custom AI models
The benefits of having personalized AI models are evident, but to carry out a proper implementation it is key to take several factors into account:
- Data quality and availability. It is necessary to have reliable, well-structured and representative data to train AI models, which implies rigorous work of collecting, cleaning and labeling the information.
- Talent and experience. Having the necessary knowledge to develop and train AI models is essential. Data scientists and AI experts play a crucial role in this process, but they are rare. Therefore, investing in talent and strategic alliances is essential.
- Infrastructure requirements. AI models demand large computing and storage capacity, so companies must have a robust, flexible and scalable IT infrastructure. This will allow them to train and deploy custom AI models efficiently, continually refining and retraining them with new data sets to keep them relevant, without the need for large investments.
“The key to success lies in having a well-defined AI strategy, which finds the right balance between data, infrastructure and knowledge. In this way, companies will be able to use AI models in a more intelligent and personalized way,” concludes Jorge Vázquez.
