K1x

Patented models read partnership statements that arrive as inconsistent documents and pull out not just headline figures but footnotes, supplemental statements and the awkward allocation items that make this work slow, reportedly in about eleven seconds per form. That is real model work. The company is also candid that it is not the whole product, stating that extraction is only step one and that the hard part is getting the right tax data to the right party. Apply the removal test and a distribution, validation and filing platform survives, operating on manually entered data, which is a painful but functioning product.

Human review is designed into the flow rather than assumed away, and one customer describes the mechanism vividly, noting that a confirmation marker on a processed item is what tells a preparer the extraction was right. Validation steps sit between extraction and distribution, and the professional who signs the return remains the accountable party throughout.

What is not described is the detail a reviewer would want: no confidence indication on an extracted value, no exception queue for documents the system could not read cleanly, and no stated threshold at which an item is routed for manual attention.

Validation and governance are named as platform capabilities and the confirmation affordance gives a preparer something concrete to check against, which is more than a black box would offer. The measurement layer is absent. No extraction accuracy figure, no error rate by form type, no evaluation methodology, no model documentation and no stated support for a firm's own review were located, which is a notable gap for output that becomes a federal filing rather than an internal analysis.

Adoption is stated at more than 40,000 organisations, and the sharpest figure is penetration rather than volume: relationships with 44 of the 100 largest institutional investors, which is a market position claim specific enough to be checked. A 175 million dollar growth investment led by a specialist software investor arrived in 2026, a large accounting firm appears as a named customer with a published case study, and multiple practitioner testimonials are attributed by role. Speed claims are quantified against a baseline, running roughly ten times faster than common tax industry software. The board includes former heads of the federal tax authority and the Treasury.

Validation and governance are described as built into the workflow rather than bolted on, and the models are proprietary and patented, trained specifically on the form types this market deals with. The unaddressed question follows directly from that: models trained on uniquely complex tax forms were trained on someone's forms, and nothing states whether customer documents contribute to that training, whether contributions are scoped per customer, or whether a firm can decline. No accuracy evidence or provider disclosure accompanies the claim.

The documents processed carry income, allocation and identifying detail for institutions and, in the family office segment, for individuals and their families, and the platform transmits that data onward to tax preparation systems, portals and the tax authorities themselves. That makes it a conduit as well as a store. No published privacy framework, retention schedule, subprocessor list or statement on how long extracted tax data persists after a filing season closes was located.

No trust centre, enumerated certification list, attestation scope or audit period was located in this pass. Certification to file electronically with the tax authority implies security requirements were met for that channel, but it is a filing authorisation rather than an information security attestation, and institutional investors and large accounting firms would have required independent reports before transmitting client tax data. The grade records what an outside buyer can verify.

K1x supplies technology and holds no licence, and it has passed one formal process worth noting: certification to transmit exempt organisation filings electronically to the federal tax authority, which is an assessed authorisation rather than an integration.

Domain grounding is specific rather than general, with the platform handling named statutory allocation provisions covering contributed property, hot assets and qualified business income, and connecting directly to federal and state taxing authorities. Governance credibility is reinforced by former heads of the tax authority and the Treasury on the board.

The subjects are documents and entities rather than people, so this reads as accuracy governance, and the stakes are unusually direct: an extracted figure does not inform a decision, it lands in a filed tax return where the taxpayer carries the penalty for an error. Customers describe accuracy as very high, which is testimony rather than measurement.

Nothing public reports an error rate, identifies which form types or allocation items the system reads least reliably, or explains how a misread value would be caught once it has flowed into a return.

Responsibility is clearly allocated by the surrounding process, since a preparer reviews and signs the return and validation steps exist to support that review, so the accountable human is never in doubt. The vendor commits to nothing beyond it.

No accuracy guarantee, no remediation term and no published error rate, and the party who bears the consequence of a misread allocation is the taxpayer facing interest and penalties on an amended return, who is usually neither the customer nor aware that extraction software was involved.

The analytical layer is described as patented and proprietary, trained on the specific form types this market handles, which tells a buyer it is owned rather than resold. The data side is inherently clean, since inputs are documents the customer already holds rather than purchased third party records, which removes a class of fourth party exposure other vendors here carry. What is not published is the rest: no model or infrastructure providers are named, no subprocessor list exists, and nothing states where tax documents are processed.

The integration reach is the product and it extends further along the chain than anything else in this index, connecting investors, advisers, accounting firms, tax preparation software, investor portals, data warehouses and the federal and state taxing authorities themselves, with certified electronic filing at the end. Data flows into returns and workpapers without re keying, which is the specific failure the market suffered from. The company also positions explicitly against vendor lock in, which is a commitment about exit rather than entry and rare enough to note.

Delivery is cloud hosted software serving domestic institutions, accounting firms and tax authorities, so cross border complexity does not arise in the form it does for global vendors here. Residency and retention still matter because tax records carry multi year retention obligations and the platform holds document sets for tens of thousands of organisations. No hosting regions, tenancy model, residency options or subprocessor chain were located.

No rates, tiers or billing unit were located. One customer testimonial refers to a price point smaller family offices can afford, which is a positioning signal rather than a number, and the buyer range from a global accounting firm to a single family office implies very different commercial shapes that nothing public distinguishes.

Coverage spans the allocator side of private markets thoroughly, addressing institutional investors, family offices, wealth managers and fund administrators, and extending to accounting firms and tax exempt organisations. Document scope covers partnership statements, information returns, wage statements and exempt organisation filings rather than one form type. Two limits hold it at B: a substantial share of the customer base is the accounting profession rather than financial institutions, and there is nothing for banks, insurers, lenders or payments firms.