AI modality: complete definition in AI for SMEs

An AI modality is a type of data an AI model can process: text, image, audio, video, code, or a combination of several (multimodal). A text model processes text; a multimodal model (GPT-4o, Gemini) processes text, images and sometimes audio in the same exchange. The modality determines what the model "sees" and what it can "produce."

What it changes for an SME

Modality opens concrete new use cases:

• Image + text: analyzing a photo of a quote to extract amounts;
• Audio + text: transcribing a meeting, extracting actions, pushing them into the CRM;
• Code + text: automating analysis script generation from a specification;
• Text alone: the majority of SME use cases — writing, classification, summarization, translation.

The trap

Confusing marketing capabilities with real capabilities. A model "accepts" an image, but processing quality varies by type: a screenshot of a table is well handled, a scanned plan less so. In fractional AI leadership, we test feasibility on your real data before promising a multimodal use case.