AI for civil-law notary practices that want more time for execution and advice
Deed preparation, Kadaster research, Wwft and UBO investigation, file building, and standard correspondence. KNB-aligned, BFT audit trail, EU-only providers. Integrates with Caesar Nx, FundamentNotariaat, NotaFlow.
Notarial practice is under pressure. The number of candidate notaries is dropping, Wwft obligations are getting heavier, the mortgage chain demands faster turnaround times, and the average purchase deed has to be executed within statutory deadlines. At the same time a large part of the working day goes into preparatory work: filling model deeds, reading Kadaster research, drafting letters to buyer and seller, documenting Wwft investigation. That work has to be done, but it no longer all has to be done by hand.
AI does not change the notary's work, but it does change the pace of preparation. A residential transfer where the factual data is already in the file gets a draft deed based on the office model, ready for the notary to walk through, correct, and make executable. A Kadaster research is summarised automatically: encumbrances, restrictions, mortgages, attachments, with direct reference to source documents. A Wwft investigation produces a draft risk assessment with UBO, sanctions lists, and adverse media. The notary reads, assesses, and signs. AI does the collection work.
DataDream works for notary practices in real estate, family, and corporate law. Notarial practice is a regulated profession. KNB rules, the duty to serve, BFT supervision, and notarial professional secrecy are hard constraints. Builds therefore use EU-only providers, with an audit trail BFT accepts, with retrieval on validated sources (Kadaster, KvK, UBO register, sanctions lists, own office positions), and with a notary in the loop on everything going out to parties, bank, or Kadaster.
Starting can be small. Often the first pilot is a draft-deed generator for standard residential transfers or a Wwft investigation assistant. One process, one team, measurable time difference per file. Only when the pilot works and the notary and staff can trust it does scale follow to mortgage, will, or corporate deeds. First lighten manual work, then larger ambitions.
Challenges
Deed preparation remains manual
Despite model deeds and standard texts in Caesar Nx or FundamentNotariaat, for every transaction a staff member has to enter factual data, check it, and bring the text in line with office style. At volume this runs to hours per file.
AI fills a draft deed based on office models and file data: buying/selling parties, cadastral designation, purchase price, transfer conditions, mortgage clauses. The notary or staff member checks and adjusts instead of typing, with a traceable audit trail of what AI proposed.
Reading Kadaster research takes a lot of time
A Kadaster extract with encumbrances, restrictions, mortgages, attachments, and parcel history is extensive. Someone has to read it, summarise it, and pull out the relevant points for the deed and the parties. At least an hour per file.
AI produces a structured summary of the Kadaster research with direct reference to source documents: which mortgages are open, which restrictions apply, which points have to be included in the transfer conditions. The notary reads the summary in minutes and checks source documents only where needed.
Wwft investigation requires disproportionate work
UBO verification, sanctions lists, PEP lists, adverse media, periodic re-assessment, documentation for the file. An hour or more per client, and at volume it adds up. When refusing service the substantiation must be watertight.
AI collects and summarises the Wwft checks, drafts a risk assessment with direct source references, and keeps an audit trail. The notary assesses personally at elevated or unclear risk, with the time to do that carefully because the collection work is done.
Standard correspondence does not type itself
Confirmation letters to buyer and seller, mortgage proposals to bank, query letters for missing documents, invitations for execution: dozens of messages per file. It takes time not currently available for the notary or candidate notary.
AI generates correspondence based on file status and office templates, in the office writing style. A staff member checks and sends. What took fifteen minutes per message now takes two minutes plus sign-off from the handler.
Finding knowledge inside your own office is hard
What is our position on that specific transfer clause? Which case law have we used before? What does the latest KNB guidance say? The information is there, but spread across files, mailboxes, and handbooks.
DataDream builds an internal knowledge base where AI searches validated sources: own office positions, KNB publications, case law via rechtspraak.nl, tax handbooks for real estate. Answers come with direct source references so the notary can click through and verify.
Results
- Draft-deed generator on office models for real estate, mortgage, will, and corporate
- Kadaster research summary with direct source references per parcel
- Wwft and UBO investigation with audit trail for BFT supervision
- Standard correspondence to buyer, seller, and bank in office style
- Internal knowledge base with KNB publications, case law, and own positions
- Anomaly detection on file documents (missing signatures, date mismatches)
- Integrates with Caesar Nx, FundamentNotariaat, NotaFlow, Hyarchis packages
- EU-only providers, no training on client data
- On-premise or private-cloud options for strict DPA requirements
- Notary in the loop on everything that goes out
Frequently asked questions
What is the difference with AI for lawyers or jurists?
A civil-law notary practice has a different workflow than a law firm. A notary works under ministerieplicht: it is not free-market service delivery, but a regulated public function with deed execution, deed registration, and archive obligation. The processes differ: deed preparation against fixed models, Kadaster research, Wwft investigation with grounds for refusal at elevated risk, file archive 100 years. DataDream builds for that specific reality instead of a generic "legal AI" that has nothing to do with notarial practice. For lawyers and jurists there is a separate page.
Does AI work with our notary software?
Yes. DataDream works with packages common in notarial practice: Caesar Nx, FundamentNotariaat, NotaFlow, and packages from Hyarchis. For Kadaster integrations the connection runs through KIK, for CDR via the notarial register. Where a package has no open integration, work happens on exports and imports (CSV, PDF, email-to-file). First priority is always that the notary and the assistant do not have to leave their work environment, otherwise AI does not produce net time savings.
Who is responsible if AI makes a mistake in a deed?
The notary remains professionally and legally responsible. AI does not change that. Execution, identification, deed content, and duty of care to parties remain fully with the notary. What AI does is speed up the preparation phase: a draft deed based on a model and the factual data, a summary of the Kadaster research, a first risk assessment for Wwft. For execution the deed sits on the desk, read, checked, and edited by hand where needed. AI does not make a deed executable, it speeds up preparing it.
Does this comply with KNB rules and BFT supervision?
KNB rules require diligence, file building, identification, and Wwft compliance. BFT supervises this. AI tooling touches these requirements in three places: (1) file building must be traceable, (2) Wwft investigation must be substantive and documented, (3) client data must remain confidential. DataDream builds with audit trail per use case (what AI proposed, what the human reviewed, what was finally recorded), EU-only providers without training on client data, and a reporting layer you can present directly to BFT supervision or a disciplinary complaint. KNB rules change; per change the setup is reassessed.
Can AI help with Wwft investigation and UBO verification?
Yes, this is one of the places where it makes an immediate difference. Civil-law notaries have heavier Wwft obligations and must refuse service at elevated risk. AI can run UBO checks against the UBO register and trade register, search sanctions lists and PEP lists, collect adverse media, and draft a first risk assessment with source references. At elevated or unclear risk the notary takes over by hand: that is where it belongs. What you gain is not "AI that decides", but a much more thorough file in less time, with an audit trail that holds up against BFT.
What about confidentiality of client data and professional secrecy?
Notarial professional secrecy is absolute. DataDream therefore works with EU-only AI providers where data is not used for training and not stored after processing. For practices with large corporate clients, mortgage chains with banks, or files with private strategic information a setup is built where data does not leave the office network (self-hosted models or EU private cloud). Per use case it is documented which data is processed and by which model. Under BFT supervision or a disciplinary complaint that is traceable.
How does such an engagement start?
First a short scan: where in the practice does most time go into preparatory work instead of execution or advice? Often that is in standard deeds (real estate transfers, mortgages, simple wills), in Kadaster research, in correspondence with buyer and seller, and in Wwft investigation. Then a focused pilot on one of those processes, for example a draft-deed generator for residential transfers or a Wwft investigation assistant. Only when the pilot works and the team values it does scale follow to wills, corporate law, or family law. No wholesale change, no "AI strategy" without grounding in practice.
Let's get acquainted.
Book a free call or send us a message. We always respond within 24 hours on business days.
Phone
+31 85 124 95 22Location
Middelburg, Zeeland
Availability
Reachable 24/7 digitally