Dono AI makes the paper trail talk.

A property has one address. The evidence behind it can be scattered across deeds, mortgages, liens and court records. Before the keys change hands, somebody must assemble that history and decide what matters. Dono AI has built its business around this peculiar gap: the house is easy to find; the story of its ownership takes work.

• The job: turn scattered property records into usable title searches, reports and commitments.
• The method: AI prepares the work; title professionals verify the output.
• The commercial bet: deliver the result inside the systems customers already use.

That last point deserves attention. A faster document reader can save minutes. A service that returns the right information to the right order, with the underlying evidence attached, can change the working day. Dono operates where those two propositions meet. Its customers include title agencies, underwriters and real estate attorneys, people for whom an attractive answer is less useful than a defensible one.

The record is only the beginning

In July 2024, the American Land Title Association reported that Dono had secured $3.7 million. It identified three founders: Tali Gross, the chief executive; Ron Likvornik, the chief technology officer; and Eyal Stern, the chief operating officer. The company, founded in 2023 according to that report, combined language models, data retrieval and machine learning to help professionals examine ownership records.

The initial proposition was specific. Collect public records about people and properties, extract the information, and let a professional interact with the resulting package. ALTA described hundreds of document formats and thousands of distinct data points. The technology was intended to support title professionals, including preliminary checks and examination. That is a more demanding assignment than summarizing a neat folder of PDFs.

The February 2026 funding announcement described four pieces of the platform: gathering records, extracting and indexing their contents, applying underwriting knowledge, and delivering the result in the customer’s required format. Those pieces can operate together or separately. This matters because a customer may already possess the documents. The expensive part of that customer’s day could be turning them into something usable.

A package goes in. A report comes out.

Dono’s report-generation service takes existing search packages and unbundles them into formatted reports or commitments. Its extraction and indexing service also supports historical records, new title plants and updates to existing ones. A title plant, despite its industrial name, is a searchable repository of property records. Maintaining one means keeping a local history usable as new documents arrive.

In its August 2026 indexing announcement, Dono described configurable extraction parameters, indexing structures and output formats that match existing database schemas. Delivery can happen through APIs, batch processing or white-label arrangements. The expertise here is partly technical and partly occupational: knowing which fields matter, how records connect and what the receiving system expects.

This places Dono between raw property data and the professional decision. Its practical alternatives include manual searching, outside abstractors, offshore processing and legacy title-plant workflows. Dono’s distinction is the combination it offers: collection, document processing, underwriting support and delivery, with human checking inside the service. Its market is professional production, where gathering evidence and assuming responsibility remain closely connected.

The best new window is the one already open

Consider the Settlor integration announced in September 2026. An agency assigns an order to Dono inside Settlor. The property and transaction information travel with it. Dono gathers and processes the records, a title professional verifies the result, and the completed search returns to the same order. The examiner reviews it and makes the call.

The business model is unusually legible in this channel: customers pay for the searches they order, with no separate platform fee. Dono’s Qualia Marketplace offering includes search, enhanced search with structured data, and examination. SoftPro Sync supports both new searches and report generation from existing packages. These integrations reduce a familiar nuisance: sending work elsewhere, then retyping the result when it comes back.

Stern’s public account of the product’s development adds a useful detail. Client conversations, he wrote, kept pointing toward transparency and control as well as speed. That helps explain the emphasis on showing the search trail. The lesson another software builder can borrow is practical: keep the customer’s working environment familiar, and make the new output easier to inspect.

Four hours comes with a human

Dono’s current website advertises an average four-hour full search and costs 45 percent below a manual process. Treat those as company-reported performance claims. The percentage describes a comparison, rather than a dollar price a buyer can put straight into a budget. A useful purchasing conversation would test the same counties, search complexity and deliverables the agency actually handles.

Local fit matters. The Rhythmic Title partnership gives North Carolina attorneys access to packages assembled from county records and the state’s VCAP and Odyssey systems. Attorneys retain visibility before issuing opinion letters. That example reveals what geographic expansion really requires: access to the right systems and outputs that fit the jurisdiction’s professional practice. A coverage count alone cannot establish either.

“The gap isn’t just technology.”Tali Gross · February 2026

There is also a boundary around the automation. In September, Dono’s governance article said its workflow does not move money, change wire instructions or grant waivers. A licensed examiner signs before delivery. For a buyer, the relevant questions extend beyond extraction accuracy to who can approve, what the system can execute and whether the evidence remains visible.

Buying the next county

In February 2026, Dono announced another $6.5 million, led by Link Ventures with lool VC and Alumni Ventures participating. Total disclosed funding reached $10.2 million. The money was earmarked for geographic expansion, deeper automation and demand beyond title insurance, including lenders, mortgage servicers and real estate investors. Funding measures investor commitment; it does not tell us what an individual search costs.

Dono’s opportunity sits in the hours between finding a record and being ready to rely on it. The company is trying to compress those hours while leaving the professional a clearer account of what happened. In a business built around ownership, that is an appealingly concrete ambition: every answer should come with a way back to the document.