A deposit arrives in a business bank account. It looks reassuring: money coming in, a balance climbing, a borrower apparently doing well. Yet the same line could record a customer payment, a transfer from another account, or the proceeds of a new loan. Three explanations, three rather different businesses. The number is the easy part. Its meaning is where the work begins.
- bluSense converts and categorizes bank-statement data for financial analysis.
- FraudLens checks whether the documents show signs of alteration.
- SBFE data adds commercial credit-payment history to the cash-flow view.
- Specialist teams handle underwriting, fraud review and compliance work alongside the software.
bluCognition has built its business in this gap between a financial record and a financial judgment. Its customers are lenders, banks and fintech companies trying to understand applicants whose evidence arrives in assorted documents and disconnected systems. The company supplies tools to organize that evidence, examine it and enrich it. It also supplies people to do the work that resists tidy automation.
The proposition has a pleasing lack of glamour. A loan application can have an elegant front end and still leave someone wrestling with a bank statement behind the scenes. bluCognition works on that second part. Its opportunity lies in the hours between receiving information and being able to trust what the information says.
The deposit with two meanings
Start with bluSense, the company’s bank-statement conversion and analytics product. It takes information that a person can read and turns it into information a system can use. The current product description includes PDF-to-JSON conversion, transaction categorization, cash-flow analysis and profit-and-loss outputs. It also describes connectivity to live banking data through Plaid and APIs for integration with lending systems.
This matters because extracting a number and classifying a number are different jobs. A statement can be transcribed perfectly while its deposits are interpreted poorly. Consider the transfer in our opening example. Counting it as sales would give a lender an inflated impression of operating activity. The practical value of categorization is that it asks the extra question before the numbers reach a decision.
At its 2023 Finovate presentation, bluCognition described transaction classification into more than 50 categories and more than 30 analytical features. Those are descriptions of the product’s scope, rather than a guarantee about any particular applicant. They do, however, suggest the ambition: make the statement a working financial record, instead of a PDF that has to be read all over again.
A PDF deserves cross-examination
There is an awkward companion problem. What if the statement itself has been altered? A conversion engine could faithfully extract an invented balance. Speed would then make the mistake arrive sooner.
FraudLens addresses document tampering. The company describes checks for inconsistent fonts, image modifications, metadata anomalies and layout irregularities. A changed amount might leave a different typeface. A supposed historical document might carry a puzzling modification history. Spacing can become an accidental witness. Office paperwork has rarely had such an interesting supporting cast.
The important distinction is between a clue and a verdict. An unusual font or recent edit can warrant attention without proving misconduct. A lender evaluating this sort of tool needs to know how flags are explained, what a reviewer can inspect and how legitimate exceptions are handled. The value is in helping the review become more selective and better informed.
bluSense and FraudLens were demonstrated together at FinovateSpring in May 2023. The pairing makes sense: one product asks what the document contains, the other examines whether its appearance and file properties deserve scrutiny. A lending workflow needs both questions answered before a neatly structured record becomes persuasive evidence.
The bank account meets the credit file
The next development changes the breadth of that evidence. On February 3, 2025, bluCognition and the Small Business Financial Exchange announced a partnership bringing bank-transaction analysis together with commercial credit-payment performance data. SBFE lists bluCognition among its additional licensees.
There are two histories here. Bank transactions show money moving through an account. Commercial payment records add information about the business’s credit obligations and repayment behavior. Neither view automatically explains the other. Bringing them together lets a lender examine current activity alongside a record of paying creditors.
“help lenders see the full picture”Sangarsh Nigam, founder and CEO, in a public LinkedIn statement
The company now calls its commercial credit offering bluConnect. Its other data product, bluIntel, enriches business profiles with public records, web presence, reviews and related information. These signals can give an analyst additional context about a business with a limited conventional credit file. They also need interpretation: a lively website is evidence of a lively website, and only one part of an assessment.
This combination places bluCognition between several familiar markets. Commercial credit bureaus supply business credit information. Bank-data services supply transaction access. Document tools extract or verify paperwork. Outsourcing firms provide reviewers. bluCognition brings elements of these jobs into one offering, with commercial lending as the organizing problem.
The SBFE relationship gives that positioning something concrete. Dun & Bradstreet, Equifax, Experian and LexisNexis Risk Solutions also appear in SBFE’s partner ecosystem. A lender’s choice depends on the job at hand: buying a credit report, building cash-flow analysis or finding a team to handle operational exceptions. bluCognition’s combination is most relevant when several of those needs arrive together.
The people behind the automation
Founder and CEO Sangarsh Nigam previously served as a senior vice president and chief credit officer for Global Commercial Payments at American Express, according to his Finovate speaker biography. Co-founder and managing director Mukesh Chamedia oversees technology development and operations from Pune, according to the company. The experience is pertinent: the products serve a workflow their leadership knows from financial services.
SANGARSH NIGAM / CEO
MUKESH CHAMEDIA / MDThe website displays customer logos including American Express, PayPal, Stripe, WEX, Fiserv, Zip, Brex and Lendio. That is the company’s presentation of its customer relationships; a logo does not specify which product a customer buys. The range nevertheless shows the intended audience: organizations with financial decisions and risk operations to manage.
The business sells software and data, managed services and consulting. Its managed work includes commercial underwriting, financial statement spreading, fraud review, identity and business checks, sanctions screening and dispute management. Consulting extends to risk strategy, model development and validation, workflow design and offshoring.
Human review is therefore part of the commercial offering. The financial-spreading service explicitly describes a semi-automated process with a human in the loop. Automation can prepare a file; an analyst can deal with its peculiarities. Selling both gives customers a way to address the queue of work as well as the software that feeds it.
Company-reported revenue, alongside bootstrapped status. A historical disclosure, not a current valuation or revenue estimate.
What a lender should put to the test
For a prospective customer, the sensible starting point is a representative set of files. Include different statement layouts, ordinary transfers, uneven income patterns and legitimate documents that look unusual. Compare the outputs with reviewed records. Measure extraction errors separately from category errors, and document flags separately from confirmed alterations. Otherwise one appealing accuracy figure can conceal several different kinds of work.
The economics also depend on that work. A buyer should compare software and data charges, integration effort and reviewer time with the existing process. Faster extraction is useful when it reduces the total handling burden; it can disappoint if uncertain categories simply create a larger review queue. That is a test of the deployment, rather than a conclusion to draw from a demonstration.
Incomplete account coverage, stale statements and weak business-context signals limit what any combined profile can say. A decision also needs a lender’s own credit policy. The company supplies evidence and operational support; the lender still has to decide how that evidence relates to its appetite for risk.
The transferable lesson is modest and useful. Separate reading the record, checking the record and judging the business. Give each step a clear purpose and a route for exceptions. bluCognition’s product mix reflects that division of labor. The bank statement can remain thoroughly boring to look at. What changes is how much of the business a lender can see through it.