Contents
20BPMator
BPMator is where you put AI to work on your documents and data in a repeatable way. Instead of running ad-hoc queries, you can configure instructions, questions, and automated steps that process incoming information consistently.
All AI operations in BPMator consume credits from your workspace balance. Outputs and findings generated by your automated processes can be retained as workspace knowledge, allowing Ask Bmator to cite them in future inquiries.
Guide Overview
Explore the following sections to set up and run your automated processes:
- Tasks: an instruction run on your documents — Use this when you need to execute a specific prompt or extraction repeatedly across sets of files.
- Questions and question lists — Use this to evaluate every incoming document against a standardized series of questions.
- Scoring and ranking subjects — Use this when you need objective scores, benchmarks, or comparative rankings derived from document answers.
- Workflows: chaining steps — Use this to link multiple actions, transformations, and decision paths into a continuous sequence.
- Data: datasets, datasources, endpoints and collectors — Use this to import external information, manage structured collections, and configure gathering mechanisms.
- Solutions: one form, one verdict — Use this when you want a simplified, prebuilt application interface that yields a direct decision or report from user input.
- Running a process from your AI assistant — Use this to trigger configured automated operations directly through conversational commands.
In this chapter
- 20.1 Tasks: an instruction run on your documents
Learn how to create and run AI tasks on documents, folders, or subjects in Bmator, configure output options, and monitor execution history and costs.
- 20.2 Questions and question lists
Learn how to create reusable questions, build ordered question lists, score documents, and export answer data in Bmator.
- 20.3 Scoring and ranking subjects
Learn how to create scoring lists, add subjects, evaluate entities against rubrics, and export ranking results in Bpmator.
- 20.4 Workflows: chaining steps
Learn how to chain tasks and collectors into multi-step workflows in Bmator, configure fan-out or fan-in execution, and run them on folders or subjects.
- 20.5 Data: datasets, datasources, endpoints and collectors
Learn how Bmator connects datasources, endpoints, collectors, and datasets to automate data collection and reporting without AI.
- 20.6 Solutions: one form, one verdict
Learn how to use and create Solutions in Bmator. Run turnkey apps with one form, get verdicts, fork Catalog solutions, or build your own with JSON.
- 20.7 Running a process from your AI assistant
Learn how to run Bmator tasks, question lists, and workflows directly from connected AI assistants like Claude using the Available as a tool setting.