Before Pramaana, I was a director of financial reporting at a mid-size company that had just gone through a regulatory examination. The examiner's focus was on a specific line in our annual report: a figure related to a deferred revenue calculation that had been the subject of an accounting policy change two years prior. They wanted to understand how that number was derived and where it could be traced in our supporting documentation.
The number was correct. I knew it was correct because I had been involved in building the model that produced it, and I had spent considerable time during the policy transition period making sure the recalculation was defensible. What I could not do, when the examiner asked, was produce the evidence chain quickly. The source documents existed. The intermediate calculations existed. The connections between them existed in my memory and in the memory of the analyst who had built the original model. But those connections were not documented in a traceable, navigable form. What should have been a 30-minute substantiation exercise became a three-week reconstruction project.
That experience is why Pramaana exists.
What the investigation revealed
During those three weeks, working through a pile of board packages, management adjustment memos, and successive versions of the deferred revenue model, I understood the problem in a way I had not when I was inside it. The issue was not that we had done bad work. The documentation practices we had used were entirely standard for a finance team of our size. The issue was that our documentation practices treated evidence as something you assembled in response to an inquiry, rather than as something you maintained continuously as a property of how the work was done.
The source documents were in the document management system. The working papers were in the audit software. The connection between them existed in the spreadsheet cells, coded by the analyst who built them, in a notation system that made perfect sense to her and required significant explanation to reconstruct for anyone else. When that analyst left the company eight months before the examination, she took the navigation key with her, not through any failure of handoff protocol, but because there was no protocol. The link between the figure and its origin was personal knowledge, not institutional documentation.
This is not an unusual situation. It is, based on conversations with finance and audit professionals over the past several years, the standard situation. The evidence exists. The connections are implicit. When they need to be explicit, the implicit ones have to be reconstructed at significant cost.
Why the problem is structural, not procedural
After the examination, I pushed for better documentation practices on the team. We added citation requirements to working paper sign-off. We created a shared drive structure that was supposed to link source documents to working papers more explicitly. We trained the team on why this mattered.
Those changes helped, marginally. But they did not solve the problem, because they addressed documentation as a separate activity from the work itself. Adding a citation step to a working paper after the fact requires the analyst to navigate back to the source document, locate the specific passage, transcribe the reference, and record it in a format that can be followed by someone else. When that additional step is performed under deadline pressure, it happens inconsistently. The citations that were most important were often the ones that got skipped, because the most material figures were also the ones requiring the most judgment, and when an analyst is working through a complex estimate, the documentation step comes last.
The structural problem is that documentation practices designed around a manual workflow will always lose to deadline pressure in edge cases. The only way to solve the problem reliably is to make the citation the default output of the work, rather than a separate step on top of it. That requires technology, not just process change.
What Devin and I set out to build
When I left that role and started talking with Devin about what became Pramaana, the core question was: what would it take to make the provenance chain a continuous, maintained artifact of working paper preparation, rather than something that exists only when someone has time to document it explicitly?
Devin had been building document processing systems for several years and had a clear view of the technical feasibility. The components that make provenance tracing possible, structured extraction from heterogeneous documents, entity recognition across financial terminology, version-aware citation linking, were mature enough to assemble into a coherent product. The question was whether the product could be built in a way that integrated with existing working paper workflows rather than requiring teams to adopt a new way of working from scratch.
That integration requirement shaped every architectural decision we made. We did not want to build a system that replaced working paper software. We wanted to build a provenance layer that sits alongside existing workflows and maintains the citation links that those workflows leave implicit. The output is not a new kind of working paper. It is the evidence chain that proves the existing working paper.
The name and what it means
Pramaana is a Sanskrit term used in logic and epistemology, referring to the means of obtaining valid knowledge. In traditional usage, it encompasses perception, inference, and testimony as the recognized sources through which a claim can be established as true. We chose it because it captures something precise about what we are building: not knowledge management in the broad sense, but the specific question of what constitutes valid evidence for a claim and whether that evidence can be traced.
In financial audit and compliance, the question is always the same: how do you know this number is right, and can you prove it? The answer requires a provenance chain, a traceable path from the assertion to the originating evidence. Pramaana is the infrastructure for building and maintaining that chain.
What we have learned since starting
Building Pramaana has confirmed the core insight from my own experience, while adding a dimension I had not fully appreciated: the problem is not limited to formal audit contexts. It shows up everywhere finance teams interact with evidence-based claims.
The most expensive manifestation is audit restatements, where a figure that was reported correctly cannot be defended because the evidence chain has degraded over time. But the same underlying problem appears in board reporting, where a CFO cannot quickly substantiate a figure that a board member questions. It appears in due diligence processes, where an acquiree's team cannot rapidly produce traceability for financial claims that an acquirer wants to verify. It appears in regulatory submissions, where the difference between a smooth examination and an extended one often comes down to how quickly the internal team can respond to evidence requests.
In all of these contexts, the issue is the same: evidence exists, the connections between evidence and claims are implicit rather than explicit, and making them explicit under pressure is expensive. The cost is not just in professional time. It is in the risk that accompanies the process of reconstruction: the possibility that a connection cannot be fully reconstructed, that the person who built it has left, that the source document was not retained in its final version, that the intermediate calculation assumed a context that no longer exists.
What we are not trying to do
It is worth being direct about what Pramaana is not. We are not building a system that replaces auditor judgment. The professional judgment required to assess whether evidence is sufficient and appropriate, whether accounting treatments are in accordance with applicable standards, whether estimates are reasonable: those remain with the audit professionals who are qualified to make them. What we are building is the infrastructure that makes exercising that judgment faster and more defensible.
We are also not trying to solve every documentation problem in financial reporting. There is a large category of figures that rest on management assumptions and market inputs rather than on directly traceable source documents. For those figures, the evidence chain is a documented rationale, not a citation link. We handle those figures differently from figures that are directly citable, and we are explicit about that distinction in how we present the system to prospective customers.
The problem we set out to solve is specific: when a figure in a working paper should be traceable to a source document, and that traceability currently requires a reconstruction exercise rather than a navigation exercise, we want to change that. The investigation I went through, the three weeks of reconstructing a chain that should have been maintained, should not be a normal cost of doing business in financial reporting. The tools to eliminate it exist. Building them into a product that fits real audit workflows is what we do.
If you work on a finance or internal audit team that has had a version of that experience, we would like to show you specifically what Pramaana does on your document set. Request a demo and we will come to you with a concrete demonstration, not a generic overview.