Whitepapers

We have not published a whitepaper yet

A whitepaper that is a sales deck with footnotes helps nobody. We will publish one when there is a question that needs more than a guide's length to answer, a method we can set out, and results we measured ourselves. Nothing we have written so far meets all three tests, so this page lists nothing. A registration form in front of an empty PDF would be worse than an empty page.

What we intend to publish

These are candidates in preparation, not commitments with dates:

• How field-level confidence should be calibrated and reported for finance documents, and why a single document score misleads.
• A matching architecture for payables: the checks, the order they run in, and the failure modes we have seen.
• GST reconciliation across many registrations: what breaks in GSTR-2B matching at volume, and what a tolerance policy should say.
• A control model for agentic AI in finance: what an agent may decide, what it must hold, and how the record is kept.

Each will state its method, say whether its figures are measured or illustrative, and name its limits. Where the data is a customer's, it will be published with that customer's consent or not at all.

What to read in the meantime

The guides already run to the length of a short paper each, and they are specific about mechanics. The automation benchmark sets out the readiness method we use and labels it as our own. The glossary defines the terms a paper would lean on, and the blog is where arguments are made before they are settled.

If you need a particular document

Procurement and security reviews often need something specific: an architecture note, answers to a questionnaire, a description of how data is handled in each deployment option. We provide those on request, under NDA where appropriate. Request a demo and say what you need, or write to connect@ptasai.com.