AI Capabilities

Creating new content that did not previously exist, across any modality

Comprehending unstructured inputs and pulling structured information out of them. The input is richer/messier than the output

Assigning labels, categories, or binary judgments to inputs. The output is a label or score, not structured data or new content

Estimating future states, values, or outcomes from historical data. The output is a numeric estimate or probability about the FUTURE

Finding relevant items from a corpus or catalog based on a query or context. The output is a ranked set of existing items or a synthesized answer grounded in retrieved sources — not newly created content or labels

Finding the best solution, allocation, or sequence under constraints. The output is a decision, plan, or configuration that maximizes or minimizes an objective

Multi-step autonomous reasoning with tool use, planning, and action. The distinguishing cognitive function is autonomous multi-step decision-making — the AI determines WHAT to do next across a sequence of steps, not just performing a single operation