Inspiration
During a routine supervisory review, a regulator receives a stack of executed syndicated loan agreements, each hundreds of pages long. Days are spent manually tracing covenant thresholds, financial definitions, exceptions, and maturity terms across dense legal text. Reviewers know that even small late-stage changes can materially alter risk, yet there is no reliable way to confirm whether the signed agreement still reflects what the bank originally approved.
When issues surface months later, they are often buried deep in legal language rather than clearly documented. The problem is not lack of effort, but lack of certainty.
Problem Statement
In the syndicated loan market, there is a gap between:
- what the bank approved, and
- what the bank signed
That gap is where undocumented risk enters the system. Loan documentation risk is not caused by the length of agreements. It arises because small, late-stage changes introduced during drafting and negotiation are effectively invisible to regulators once the agreement is executed. Regulators typically receive syndicated loan agreements only after execution, as static PDFs, with no reliable way to verify whether the final legal terms still conform to approved credit conditions.
Regulators do not need additional monitoring systems. They need certainty that the executed loan matches what was approved.
Existing Solutions (and why they fall short)
Today’s loan documentation tools support drafting, comparison, and administration, but they are not built for regulatory verification at execution. Drafting platforms focus on negotiation, comparison tools show text changes without highlighting risk impact, and loan systems rely on manually entered data. As a result, regulators still review long executed agreements manually, with no reliable way to confirm that signed terms match what was approved.
The Gap
Banks approve loans based on structured credit papers, while regulators review loans using executed legal agreements. Between approval and execution, documents are revised under time pressure, and legally binding changes to covenants, definitions, and thresholds may occur without being clearly traced back to the original approval. Existing tools focus either before signing or after monitoring, leaving the execution moment—when risk becomes fixed—largely ungoverned.
Solution
DocConform is an execution-stage verification tool designed for regulatory review. It verifies that executed loan agreements conform to the terms banks approved and surfaces deviations regulators actually care about.
Rather than monitoring loans or generating new reports, DocConform reconstructs, from the executed agreement itself, the legally binding economic terms supervisors already examine and validates them against approved credit conditions. Where discrepancies, inconsistencies, or undocumented changes exist, they are made explicit before supervision begins.
DocConform does not create new regulatory requirements. It makes existing supervisory review faster, clearer, and more reliable.
What DocConform does
DocConform treats the executed loan agreement as the sole source of legal truth.
It identifies the specific approval-level terms regulators routinely examine, including:
- facility amount and currency
- maturity date
- benchmark and margin
- covenant thresholds and test mechanics
- key financial definitions and exceptions
- presence of required clauses
These terms are traced across the document and linked directly to their source text. Where terms appear multiple times, are defined inconsistently, or diverge from approved credit conditions, those differences are surfaced explicitly rather than remaining embedded in legal language. link The output is a structured, regulator-ready view of the executed agreement, supported by evidence and an audit trail.
Why DocConform is Different
DocConform is not a document digitisation tool. It is a governance tool.
It does not aim to summarise documents, monitor performance, or predict risk. Instead, it answers a narrow but critical supervisory question:
Does the executed agreement reflect what the bank approved?
By tying extracted terms directly to their legal source and comparing them to approved conditions, DocConform turns regulatory review from a document-reading exercise into a verification exercise.
Target Market
DocConform is designed for regulatory-facing loan review, not day-to-day loan administration.
Primary users
- Banking supervisors conducting credit file inspections
- Regulatory review teams within central banks and prudential authorities
Secondary users
- Bank risk and compliance teams preparing for supervisory reviews
- Legal and loan operations teams responsible for document integrity
DocConform is not intended for traders, portfolio managers, or real-time monitoring teams. It supports periodic, evidence-based supervisory review, where regulators experience the greatest friction today.
Why Now
Syndicated loan agreements are becoming longer and more bespoke, while supervisory expectations around governance, auditability, and evidence continue to increase. At the same time, regulators face growing resource constraints.
DocConform does not ask regulators to change how they supervise. It gives them the clarity and certainty they already seek, without adding new processes or reporting burdens.
How we built it
DocConform is a full-stack web app built around a simple demo flow (upload → verify → evidence → export), designed for supervisory constraints: traceability and auditability.
Tech Stack
Frontend
React + Vite + TypeScript
Tailwind CSS
API base via VITE_API_URL
Backend
Python + Django
Django REST Framework (DRF)
Multipart file upload (multipart/form-data) for executed agreement + approved summary
Document Processing Deterministic PDF text extraction (page-by-page) Evidence-first term extraction (anchored rules + pattern checks for dates, amounts, bps, ratios) Evidence stored per term/issue (snippet + page reference or char-offset range)
AI (Used Carefully) Normalization of extracted values (e.g., percent → bps, date formats → ISO) AI never invents terms; outputs must link back to evidence text
Integrity, Audit, Exports SHA-256 hashes for uploaded files (tamper-evident) Audit trail events: upload → extract → validate → export Exports: JSON (metadata + hashes + terms + issues + audit), CSV (side-by-side term comparison)
Deployment Frontend: Vercel Backend: Render CORS configured for local + deployed domains
Value proposition
For regulators
- Confirms that executed loan agreements conform to approved credit terms
- Reduces manual document review while increasing supervisory certainty
- Makes late-stage deviations visible before examinations begin
For banks
- Identifies undocumented risk before regulatory review
- Reduces supervisory findings and costly remediation
- Strengthens trust with regulators without changing existing workflows
For the market
- Improves confidence in syndicated loan documentation
- Addresses document-level risk without adding monitoring systems
Scalability Potential
- Applies consistently across syndicated loans regardless of size, sector, or jurisdiction
- Can be adopted incrementally by banks or supervisors without changing existing documentation workflows
- Scales across portfolios by verifying documents one-by-one rather than requiring market-wide coordination
- Supports expansion to additional document types (e.g. amendments, waivers, refinancings) using the same validation approach
- Enables standardised supervisory review without requiring standardised loan documentation
Potential Efficiency Gains
- Reduces time spent by regulators manually locating and cross-checking key terms across long loan agreements
- Lowers the effort required to verify alignment between approved credit terms and executed documentation
- Decreases the likelihood of repeat reviews caused by late discovery of undocumented deviations
- Simplifies preparation for supervisory reviews by producing a clear, structured view of executed agreements
- Enables banks to address documentation issues proactively rather than during regulatory examinations
Potential Impact
- Improves supervisory confidence in the accuracy of executed loan documentation
- Reduces the introduction of undocumented risk into regulated loan portfolios
- Lowers the frequency and severity of supervisory findings related to documentation issues
- Strengthens trust between banks and regulators through clearer, verifiable credit records
- Contributes to more consistent and reliable oversight of the syndicated loan market
Challenges We Ran into
- Designing a verification system that operates on executed documents without altering legal meaning
- Identifying and tracing cross-referenced terms consistently across long, complex agreements
- Balancing structured validation with the need to avoid legal interpretation or judgment
- Ensuring the solution supports regulatory review without duplicating existing monitoring tools
- Making the output clear and actionable for supervisors while remaining flexible across loan structures
- Many of these challenges required revisiting early assumptions and refining both the scope and design of the system
Accomplishments that We are Proud of
- Clearly identifying a real supervisory gap between approved credit terms and executed loan documentation
- Designing a solution focused on regulatory certainty rather than additional monitoring or automation
- Building a working prototype that verifies executed loan agreements against approved terms
- Aligning the product design with actual regulatory review practices and constraints
- Creating a clear, judge-friendly interface that demonstrates the value without disrupting existing workflows
What We Learned
- Regulatory certainty depends more on document integrity than on additional monitoring
- Small documentation changes can introduce material risk if they are not made explicit
- Structured validation is most effective when it supports, rather than replaces, human judgment
- Solutions gain credibility when they align with existing regulatory and legal workflows
- Iteration and scope discipline are essential when working in highly regulated environments
What's Next for DocConform
- Extending validation beyond base agreements to amendments, waivers, and refinancing documents
- Supporting portfolio-level review across multiple executed agreements
- Refining validation rules through feedback from regulators and loan market practitioners
- Enhancing export and reporting formats to align with supervisory review processes
- Exploring integration with existing document management and credit systems
Log in or sign up for Devpost to join the conversation.