Black Ore

Tax preparation is one of the few professional workflows where the case for automation writes itself. The work is seasonal, brutally compressed, highly repetitive, and staffed by people who are expensive and increasingly hard to hire.

Black Ore is built for that specific pain, and it is built for the firms that feel it hardest. Tax Autopilot is not a chatbot bolted onto a tax product; it is positioned as an end-to-end system that gathers client documents, classifies them, extracts the data, performs the calculations, assembles the workpapers, flags anomalies for a reviewer and produces an audit-ready package.

What follows covers what the four modules actually do, why the integration list matters more than any feature, how to read the efficiency claims on the homepage, and the questions a firm should get answered before signing.

Black Ore At Firm Level

The Short Version For Firm Leadership

• Tax Autopilot covers four stages — admin, preparation, review and audit support — rather than automating one step and leaving the rest manual.
• It works inside your existing tax software, with stated integrations for CCH Axcess, ProSystem fx, UltraTax CS, GoSystem Tax RS, Lacerte, ProConnect, ProSeries, Drake and ATX.
• Individual returns are the live product. Entity and fiduciary forms are presented as arriving rather than shipped, which sets the realistic scope of a first engagement.
• The customer list is the strongest evidence: Baker Tilly, Withum and Grassi among Top 25 firms, with Honkamp and Maxwell Locke & Ritter in the Top 100.
• There is no self-serve entry. Access runs through a demo and a sales conversation, so budget and procurement time are part of the cost.

Four Stages, Not One Clever Trick

Most AI tax tools attack a single step — usually document extraction, because it is the easiest to demonstrate; research assistants such as TaxGPT never touch return production at all. Black Ore's structure is broader, and the breadth is the point.

The administrative module generates client data requests, classifies documents as they arrive and assembles the binder and workpapers without someone naming and filing each PDF. Anyone who has run a tax season knows this stage consumes an unreasonable share of junior hours for work that produces no professional judgement.

Preparation extracts figures from source documents, runs the calculations and pushes the result into the firm's tax engine, producing an Excel tie-out package alongside. Review applies anomaly and error detection, gives the reviewer a copilot over the return, and packages it for signoff. The audit module keeps documents indexed with evidence linked to the steps that used them, and exports the package.

The reason the four-stage shape matters is that automating extraction alone rarely changes the economics. If the data lands accurately but a person still has to assemble the binder, chase the missing 1099 and re-key the result into CCH, the saving is a fraction of what the demo implied.

Automating one step of a return changes a task. Automating the handoffs between steps is what changes a season.

Which Returns It Actually Handles Today

Form coverage is the first question to ask, and the site answers it plainly enough: individual returns are the live capability, with fiduciary and partnership forms and the associated K-1 and K-3 schedules presented as arriving rather than shipped.

That is a sensible sequencing decision. Individual returns are the highest-volume, most templated work in most firms, which makes them both the best automation target and the safest place to prove a system.

For a buyer it sets the realistic shape of a first year. If your firm's pain is complex partnership allocations, this is a roadmap conversation rather than a purchase, and analysis tools such as Blue J address a different part of that problem. If your pain is a wall of individual returns between February and April, the live product addresses it. Confirm the current status of each form family in writing before committing — roadmap language moves faster than roadmaps do.

Nine Tax Engines, And Why That List Matters Most

Black Ore states integrations with ATX, CCH Axcess Tax, CCH ProSystem fx, Drake Tax, GoSystem Tax RS, Intuit Lacerte, Intuit ProConnect, Intuit ProSeries and UltraTax CS.

That list is the most commercially serious thing on the site. Tax software is the deepest incumbency in professional services — a firm's engine choice is entangled with its billing, its document management, its staff training and a decade of prior-year returns. A tool that requires migrating off CCH Axcess will not be bought, however good it is.

Nine engines also implies real engineering investment rather than a single connector demonstrated on stage. Each of these products has its own data model, its own import mechanics and its own quirks across versions.

The question to press in a demo is what depth means for your engine specifically. Read-only import of extracted figures and full round-trip write-back with diagnostics are both truthfully described as an integration, and they are entirely different products in practice.

• Ask which of the nine integrations your firm's engine uses, and which version it was certified against.
• Ask whether data flows one way or round-trips, and what happens when the engine rejects a value.
• Ask how prior-year data is used, since carryforward is where most extraction tools quietly break.
• Ask what the reviewer sees when the system is unsure — a flagged field is useful, a silently confident wrong number is not.
• Ask who is accountable for an error that reaches a filed return, and get the answer from the engagement letter rather than the sales deck.

Reading The Efficiency Numbers Honestly

The homepage carries a set of striking figures: better than 99% accuracy on data extraction, 98% time savings on preparation, 80% lower cost per return, more than double the review speed, and a claim of zero turnaround time.

These deserve scrutiny, and the last one deserves the most. Zero turnaround time is not a measurable quantity — every process has a duration. It reads as marketing emphasis rather than a metric, and its presence alongside the others is a reason to treat the whole set as directional.

The 98% figure is the one to interrogate in a demo. Time savings against what baseline? A return prepared from scratch by a first-year associate with no prior-year file is a very different denominator from a returning client's straightforward filing. Firms measuring their own before-and-after almost always find the honest number is large, real, and considerably below the marketing figure.

None of this means the product does not work. It means the evidence you should weigh is the customer list, not the percentages. Baker Tilly, Withum and Grassi are firms with procurement functions, security reviews and professional liability exposure; they do not deploy a tax system on the strength of a homepage claim. That is a far stronger signal than any number the vendor chose to print.

What A Firm Should Verify Before Signing

Because there is no self-serve tier, evaluation happens inside a sales process. That makes a written checklist more valuable than usual — it is easy to leave a good demo without having asked the questions that matter in April.

Run this sequence rather than reacting to the demo script.

1. Bring your own returns. Ask to run a sample of real prior-year files, including two you found genuinely difficult, rather than watching a prepared demonstration.
2. Check the exception path first. Ask to see what happens on a return the system cannot complete, because that is the case that will consume your staff's time.
3. Get the security review started early. Ask for the SOC 2 report under NDA, plus data residency, retention, deletion and subprocessor detail — not the summary claim on the site.
4. Confirm form coverage and dates in writing for every form family your firm files in volume.
5. Pilot on a defined slice of the season with a measured baseline, so your own before-and-after number replaces the vendor's.

Security, And Why It Is The Gating Question Here

A tax platform holds the most sensitive dataset a household has: identifiers, income, dependants, brokerage activity, health-related deductions and the documents behind all of it. For a firm, a breach here is not an IT incident but a professional and regulatory one.

Black Ore states US-based, SOC 2-compliant infrastructure. That is the right claim to make and it is the beginning of the conversation rather than the end of it. A firm should ask which SOC 2 type the report covers, what period it spans, which trust criteria are in scope, and what the exceptions were.

The site publishes its privacy policy and terms of service, which are the documents that actually bind. Read the data-use provisions closely, particularly anything covering model improvement, and get a written answer on whether client data is ever used to develop the product.

For a firm with an information security programme, none of this is unusual diligence — it is the same review a law firm would run before deploying Harvey AI. It is worth stating only because AI vendors are frequently newer than the firms buying them, and the depth of a security programme rarely matches the confidence of a homepage.

• Ask for the SOC 2 report itself, its type, and its observation period.
• Get an explicit written position on whether client data trains or improves models.
• Confirm data residency and the retention and deletion schedule for source documents.
• Request the subprocessor list, including which model providers process return data.
• Agree breach notification timelines that match your own client commitments.

How Black Ore Is Sold

There is no published price, no free tier and no trial. Access begins with a demo request, which is the standard motion for enterprise software sold to firms with procurement functions.

That structure carries real information. Contact-sales pricing usually means the figure is negotiated against firm size, return volume, seat count and term, and that the vendor expects an implementation alongside the licence. It also means the true cost includes your own people's time: security review, integration work, staff training and a supervised first season.

The investor group — Oak HC/FT, Andreessen Horowitz, Founders Fund, General Catalyst, Khosla Ventures and Trust Ventures — tells you the company is funded to sell this way and is not under pressure to convert self-serve signups. For a buyer, well-funded is mostly good news: the vendor will still exist next season.

Budget for the evaluation as well as the licence. A pilot that measures a real baseline against a real slice of your season is the only way to know what the system is worth to your firm, and it is not free to run.