Contextual Intelligence
HCI Research • Ameer Omidvar
Definition
AI can not strive to know everything; it should understand what matters. True intelligence is contextual—rooted in the environment, data, and priorities of a specific domain or individual.
Purpose
To move from general knowledge toward embedded understanding. By training AI on specialized or proprietary data, it becomes biased toward relevance, reflecting the culture, memory, and goals of its users.
Design Guidance
Treat data as perspective: your content defines what “intelligence” means in your context.
Favor depth over breadth—a smaller, domain-specific model often outperforms a general one.
Allow adaptive bias: users or organizations should be able to tune what the AI values.
Integrate memory—AI should accumulate experience, not just recall facts.
Prioritize trust: users relate more to an AI that understands their world than one that knows all worlds.
Example
A law firm’s AI assistant trained on the firm’s own cases, templates, and workflows can answer questions, draft documents, and even recall office-specific details far better than a general legal chatbot.