AI solutionsShared across all subject areas

Multimodal Instruction Processing

Executes tasks based on instructions combining text and visual context (e.g., 'extract the red box content').

Description

Accepts an instruction that refers to something visible — a region, a highlighted section, a marked-up page — and acts on it. The model must resolve the reference before it can execute, which is a materially harder problem than either understanding text or reading an image alone.

When it fits

A person is working with a document and wants to point rather than describe. Exception investigation and document review, where the useful instruction is 'explain this bit'.

When it does not fit

Batch processing. If no human is present to point, there is no instruction to resolve and simpler extraction is the right tool.

Governance requirement

The system should state what it understood the reference to mean before acting, particularly where the action is not trivially reversible.

Characteristic failure

Referential ambiguity resolved confidently and wrongly. 'The total at the bottom' has several candidates on a page with subtotals, and the model rarely says it was unsure.

Example

An accountant investigating a variance highlights a section of a settlement report and asks what the deduction relates to, rather than describing where it sits on the page.

AI solution components11
  • Instruction Understanding (Text + Image)
  • Multimodal Input Parser
  • Intent & Context Resolution
  • LLM-Based Reasoning Engine
  • Image-Guided Prompt Expansion
  • LLM Tool Invocation & Planning
  • Orchestration Layer for Execution
  • Fallback & Escalation Strategy
  • Prompt Diagnostics Layer
  • User-Guided Clarification Loop
  • Visual Grounding Score