KuCept Solutions, a joint venture between real estate law firm KUCERA Rechtsanwälte and Darmstadt-based company Spryfox, is bringing Legal_ALAN to market, a specialised platform for software-supported analysis of real estate documents. The system systematically prepares datasets using machine learning and consolidates them in a searchable environment. The service is aimed at institutional investors, asset managers and law firms wishing to review document inventories on a self-service basis and evaluate transaction-relevant content.
At launch, the solution is tailored to retail properties and retail parks. Extensions for residential properties and other asset classes are in preparation. In addition, the application is intended to serve further user groups in future, such as project developers or sellers conducting vendor due diligence.
Automated structuring of data rooms and contracts
The platform aims for precise structuring of complex documentation. Specific contractual parameters can be systematically identified, extracted and compared:
- —Terms and contractual option rights in leases
- —Indexation clauses and separate contractual arrangements
- —Cross-checking of inventories against document lists to identify missing addenda
- —Detection of conflicting information within the data room
The platform does not provide legal advice; the individual legal assessment remains with users and their legal counsel. Dr Stefan Kucera, founding partner of KUCERA Rechtsanwälte and managing director of KuCept Solutions, explained that artificial intelligence had arrived in the real estate sector and was becoming a central tool. However, the technology is not seen as a replacement for legal reviews, but rather as a complementary instrument. By preparing review processes more efficiently, more room remains for the assessment, structuring and negotiation of transactions.
Process-controlled auditability and European hosting
The platform incorporates more than two decades of transaction experience from the participating law firm alongside the expertise of Spryfox data scientists. Unlike open-ended language models, the system operates along a defined workflow: documents are assigned to fixed categories and recorded via uniform attribute schemas and asset-specific audit catalogues. Because every entry remains directly linked to its source reference in the original document, investors, financing institutions and investment committees can verify data provenance at any time.
Technologically, the solution combines language models, optical character recognition and semantic analysis. Johannes Wowra, Head of AI Engineering & Products at Spryfox, stressed that the infrastructure automatically reads and auditably organises thousands of documents. The software runs exclusively on hosted instances in European data centres. Additional interfaces with existing data rooms and ERP systems are planned for future development stages.
