Large language model (LLM)

A language engine that reads and writes text. It can draft, summarize, answer questions, or extract information, but it can still be wrong.

Where it creates value

Best suited to work like this.

Drafting, summarization, information extraction, conversational assistance, classification, and code support where outputs can be reviewed.

What it does

The intelligent task.

Generate, transform, classify, summarize, extract, and reason over natural-language and code inputs.

Typical value

Reduces effort in drafting, summarizing, extracting, and navigating language-heavy work while keeping people focused on judgment and approval.

Where it performs well: Flexible across many language tasks and can work with limited task-specific software development.

Value maturity · editorial assessment Scaling

Value has moved beyond isolated pilots and is being scaled across knowledge-intensive work, although reliable enterprise operation still depends on integration, governance, and human review.

Emerging
Demonstrated
Scaling
Established

Back

TEC-0001 · How it works in practice

Large language model (LLM)

Read the card type together with the use case. A component can be central in one solution and enabling in another.

01 · Inputs

Prompts, documents, structured context, instructions, examples, and optional tool results.

02 · Intelligent task

Generate, transform, classify, summarize, extract, and reason over natural-language and code inputs.

03 · Outputs

Generated text, structured data, classifications, code, explanations, or tool requests.

Common application patterns

Drafting
summarization
information extraction
conversational assistance
classification
and code support where outputs can be reviewed

Enablers

What must be in place.

A clearly defined task, suitable model, prompt and context design, evaluation set, access controls, monitoring, and a human fallback.

Where it struggles

Limitations and drawbacks.

Can hallucinate, omit context, reproduce bias, vary between runs, and present uncertain answers confidently.

Human role

People remain accountable.

People define the task and limits, review material outputs, resolve uncertainty, and remain accountable for decisions.

Connected use cases

Examples from the current use-case library.

Find safety signals earlierDetect adverse event signals in literatureSafety · Pilot Reduce document workloadAutomate regulatory document preparationRegulatory · Pilot

Continue exploring

The card is the simple front door. The website can hold the full system behind it.

Use the technology record to connect applications, processes, enablers, risks, controls, evidence, and implementation experience without placing all of it on the printable card.

Browse technologies