Most B2C leaders are captivated by developing the newest generative AI tools or agentic solutions. But since AI models are only as good as the data they learn from, the real differentiator lies in the inputs — specifically, unstructured data.

GenAI thrives on the messy stuff — images, video, speech, behavioral signals — and it’s transforming how brands operate, engage, and grow. Yet 70% of AI decision-makers still believe that predictive models can only train on structured data; this is a costly misconception.

Structured Data Got Us Here — Unstructured Data Will Take Us Further

Structured data powered machine-learning models for decades. But it’s no longer enough. Unstructured data, previously seldom used due to tech limitations, is now the fuel for unlocking more advanced automating capabilities. Think customer reviews, call transcripts, social posts, and facial patterns — all now usable thanks to AI. These are three unstructured data sources you can’t ignore: 

  1. Computer vision. Computer vision (CV) enables machines to interpret and understand visual information (images, videos, live camera feeds, etc.) in a way that mimics human sight. CV innovations in camera capabilities are turning consumer smartphones into guided buying tools, and visual data is transforming physical spaces into insight engines. All of this is leading to more relevance and value — in other words, personalization — across the customer journey. Jordan’s Furniture, for example, uses multiple sensors to scan an in-store customer’s body dimensions, distributions, and weight to find that individual the “perfect mattress.”
  2. Natural language processing (NLP). Natural language processing allows machines to interpret and respond to human language, both written and spoken, bridging the gap between raw text and business strategy. NLP turns reviews, chats, and call transcripts into actionable insights. JPMorganChase, for example, has developed IndexGPT, which uses NLP to curate thematic investment baskets.
  3. Biometrics. While biometrics is just beginning to make its way into purchasing touchpoints, consumers are already using it to unlock their phones, start cars, secure laptops, and move through the airport. Since biometrics can be used to personalize the customer experience while improving security, it makes sense that businesses want to expand its use cases. L’Oréal recently announced launching an in-store device called Cell BioPrint that does personalized skin analysis and provides recommendations to shoppers.

Consumers expect brands to “get” them. But only 39% of US online adults say they feel “surprised and delighted” when they shop. Unstructured data, when captured and used effectively, helps brands meet these rising expectations, automate smarter, and differentiate in moments that matter.

How Forrester Can Help

To learn more, read Invest In Unstructured Data Inputs To Improve GenAI Outputs. For a deeper dive, Forrester clients can schedule a guidance session with me.

In the meantime, please visit my profile page and click “Follow” to be notified whenever I publish research.