Use AI to prepare information for the people and systems that need it. FormKiQ's AI Processing and Analysis module applies configured models and prompts to documents, then connects the results to metadata, search, review, and business workflows.
In a customer-managed deployment, model requests use Amazon Bedrock through your AWS account and permissions. You control the primary document and metadata resources; the model and inference-profile configuration determine where inference is processed.
Explore an intake exampleFrom incoming document to reviewed information
- 1 · Define the task
Choose what to extract
Select a document type and define the information the next stage needs, such as an application reference, applicant name, or agreement date. Use OCR when the input requires text recognition.
Result: a configured processing task for the document type.
- 2 · Structure the result
Map values to your records
Use extraction results, mappings, and validation rules to populate document attributes. Keep the original content available so reviewers can inspect the source.
Result: information that the application can search, validate, and route.
- 3 · Apply the review policy
Resolve exceptions before use
Configure the workflow to send documents to the appropriate review queue and capture an approve or reject decision before continuing.
Result: reviewed information ready for the next business step.
Illustrative example · fictional content and results
An application with a missing reference
Source document excerpt
Applicant: Sample Orchard Company
Equipment: Packaging workstation
Requested term: 24 months
Application reference: —
The review process requires an application reference before agreement preparation.
Fields to review
| Field | Illustrative result |
|---|---|
| Applicant | Sample Orchard Company |
| Equipment | Packaging workstation |
| Term | 24 months |
| Application reference | Missing — review required |
The reviewer resolves the missing reference against the application record, then approves the document for the next step.
This example demonstrates a configured validation and review flow. It is not an AI accuracy measurement or a live model response. The fields, validation rules, and review routing are selected for the implementation.
Choose the task that supports the workflow
| Task | Useful outcome | What to configure |
|---|---|---|
| Classification | Organise incoming documents by relevant business type. | Document categories, prompts, and handling of ambiguous results. |
| Extraction | Make selected values available to search and downstream applications. | Required fields, mappings, validation, and exception handling. |
| Summarisation and analysis | Prepare a synopsis or structured analysis for a reviewer. | Questions, output structure, source access, and review requirements. |
| Draft content | Prepare text for a letter, response, or other document. | Approved context and human review; Document Generation supplies repeatable document formatting. |
Control the document and processing boundaries
| Boundary | What you control |
|---|---|
| Documents and metadata | Primary storage, access policies, and operating controls in your AWS account. |
| Model invocation | The models your deployment can access through Amazon Bedrock and the permissions used to invoke them. |
| Inference geography | The model and selected inference profile. Geographic profiles route within their defined geography; global profiles can route across supported AWS Regions worldwide. |
| Use of results | Prompts, mappings, validation, workflow steps, and the decisions that require human review. |
| External integrations | Which connected systems receive documents or resulting data, under their own access and processing terms. |
These controls help align processing with a residency and sovereignty strategy. AWS documents inference-profile geography. For operational support, compare vendor-managed and hybrid deployments with the customer-managed model.
When FormKiQ fits
Your document workflow needs AI alongside customer data ownership, API integration, governed records, and review steps. Packaged Business Solutions can supply the application context, including financial applications, contracts, and records.
When another option may fit
A specialist extraction service may be a more direct choice when a single document type and its measured extraction performance are the entire requirement. If documents already live in a platform that meets your governance needs, its integrated AI may require less implementation work.
Scope the module and the application together
AI Processing and Analysis is an add-on for Advanced and Enterprise. The implementation selects the models, prompts, document types, mappings, review steps, and integration requirements. Base platform licensing, modules, implementation, and AWS model and infrastructure usage are separate parts of the total cost.
OCR, rule-based mappings, foundation-model analysis, and retrieval through KnowledgeBase serve different purposes. A solution can combine them, but selecting one does not mean that every capability is included.
Editions and pricing · How mappings work · Review queues and workflow decisions · OCR and classification architecture
Discuss one representative intake workflow
Bring the document types, fields, review decisions, and processing-location requirements that matter. We can identify the appropriate modules and a scoped evaluation using representative examples.
Book an AI document workflow review