# Process quantification

## Why it is hard

Enterprise processes live in PDFs, slides, tribal handoffs, and half-updated SOPs. Making them **formal, drillable, and measurable** is complex and error-prone:

- Ambiguity: the same step means different things to different roles.
- Branching reality: real work has exceptions, handoffs, and failure modes — not a single happy path.
- Evidence: knowing who can *do* the process (not only who signed the document) requires behavior, not just text.
- Change: when markets or rules shift, static documents lag; unverified updates create silent risk.

## Where generative AI is weak

Raw generative AI is strong at **process-like text** (summaries, rewrite, plausible SOP prose). It is still weak at producing **reliable formal structure** that is:

- drillable end-to-end as a scenario people can run,
- measurable for decision behavior and outcomes,
- stable enough for training, ops, and risk review.

Producing fluent text about a process is not the same as quantifying and stress-testing the process.

## What SpeedL does

**SpeedL** (product/service) focuses on turning documents and tacit knowledge into **clear, drillable processes** with trackable decision behavior and analyzable outcomes — for enterprise process, training, and operations.

Typical flow (product narrative, not a claim of automated perfection):

1. Upload enterprise documents (PDF, Word, PPT, etc.).
2. AI-assisted scenario construction with human collaboration.
3. People run drills; paths and choices are recorded for insight before real loss.

## Doc2Sim (core technology)

**Doc2Sim** is the core process-quantification technology behind SpeedL: docs/SOPs → drillable, quantifiable process and decision scenarios. Mention it as technology, not as a separate marketing brand equal to SpeedL.

SpeedL is **not** defined as a decision-tree tool. Tree/graph structures may be an implementation subset.

## Related

- [What is SpeedL](https://speedl.ai/docs/what-is-speedl.md)
- [Use cases](https://speedl.ai/docs/use-cases.md)
- [FAQ](https://speedl.ai/docs/faq.md)
- Contact / Demo: https://speedl.ai/contact