> ## Documentation Index
> Fetch the complete documentation index at: https://sensible.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Choosing an extraction approach

As you scale up and encounter document complexity, you can optimize extraction performance by choosing between LLMs or layout-based extraction methods.

See the following diagram for an overview of when to use Sensible Instruct (LLMs) or SenseML (layout-based) for extractions:

![Click to enlarge](https://mintlify.s3-us-west-1.amazonaws.com/sensible/assets/v0/images/screenshots/document_landscape.png)

Sensible recommends using large language model (LLM) prompts for free-form, highly variable documents, and layout-based, or "rule-based" queries for structured, less-variable documents. You can combine both strategies since they're fully compatible with each other.

See the following table to learn more about extraction strategies:

| extraction method | notes                                                                                                                                                                                                                                                                                                                                                                                                                    | get started                                                                                     |
| ----------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------- |
| LLMs              | You describe the document data that you want to extract, using Sensible's visual authoring tool, *Sensible Instruct*. Sensible uses GPT-4 and other large language models (LLMs) to extract data from your documents.                                                                                                                                                                                                    | [Getting started](/welcome/getting-started-ai)                                                  |
| layout-based      | To extract from complex document layouts, use *SenseML*, a superset of Sensible Instruct. SenseML is a JSON-formatted query language that combines layout-based queries with LLM-based queries. When either strategy can work, Sensible recommends layout-based queries for the sake of fast performance and deterministic output.                                                                                       | [Getting started with layout-based extractions](/welcome/getting-started-ai)                    |
| out-of-the-box    | Sensible provides out-of-the-box extraction configurations for common business and tax forms. Use Sensible's pre-built, open-source [configuration library](https://github.com/sensible-hq/sensible-configuration-library/) to extract key information from tax forms such as 1099s, major carrier insurance declaration pages, and other documents. Then tweak the pre-built configurations for your custom data needs. | [Getting started with out-of-the-box extractions](/layout-based-extractions/library-quickstart) |

See the following table for an overview of the pros and cons of LLMs versus layout-based extraction:

|                                                              | LLM (Sensible Instruct)                                                                                                                                                                                 | layout-based (SenseML)                                                                                                                                                               |
| ------------------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| Technical expertise required                                 | For nontechnical users. Describe what you want to extract in a prompt to an LLM. For example, "the policy period" or "total amount invoiced".                                                           | Offers highly configurable JSON-based extraction configuration for technical users. For example, write instructions in JSON to grab the second cell in a column headed by "premium." |
| Workflow automation                                          | Suited to workflows that include human review or that are fault-tolerant.                                                                                                                               | Suited to automated workflows that require predictable results and validation.                                                                                                       |
| Document variability                                         | Suited to documents that are unstructured or that have a large number of layout variations or revisions.                                                                                                | Suited to structured documents with a finite number of variations, where you know the layout of the document in advance.                                                             |
| Deterministic                                                | No                                                                                                                                                                                                      | Yes. Find the information in the document using anchoring text and layout data.                                                                                                      |
| Handles repeating layouts                                    | Use [List](/llm-based-extractions/prompt-tips/list-tips) method.                                                                                                                                        | Use [sections](/senseml-reference/sections/index-sections) for highly complex repeating substructures, for example, [loss runs](/senseml-reference/sections/index-sections).         |
| Handles non-text images (photos, illustrations, charts, etc) | To extract data about images ("is the building in this picture multistory?", use [Query Group](/senseml-reference/llm-based-methods/query-group) method with the Multimodal Engine parameter configured | No                                                                                                                                                                                   |
| Performance                                                  | Data extraction takes a few seconds for each Instruct method.                                                                                                                                           | Offers faster performance in general. For more information, see [Optimizing extraction performance](/best-practices/tutorials/performance).                                          |