BotMan AI has built a migration pipeline that maps code, data and business rules of legacy applications into a knowledge graph before rebuilding them. Founder and chief technology officer Alejandro D'Andrea described the approach in an interview with John Furrier for theCUBE + NYSE Wired: AI Luminaries series. The system ingests old code bases such as Delphi, Visual Basic and Oracle Forms and rewrites them into a new architecture. For companies stuck on outdated platforms, this points to faster modernization without manual rediscovery of logic.
How BotMan rebuilds legacy architecture
According to D'Andrea, the pipeline does not translate old code line by line into a newer language. Instead it first builds a representation of how components, dependencies, data models and business rules relate to one another. That map becomes the working context for agents that carry out the rebuild. The company says it ingests the whole application at once and migrates it fully autonomously. Graphs play a central part in holding that system-wide view during the process.
The logic behind the design is that model capability is no longer the constraint. D'Andrea said intelligence is not the problem anymore, while context remains the bottleneck for agents doing a particular task. Without background, each task requires a long explanation of what the system does and how parts connect. A knowledge graph prepared in advance gives the agent that background automatically. In his formulation, the agent already knows much of what the team knows.
The same mechanism extends beyond code to documents and operational information. D'Andrea gave the example of an incoming legal document that is linked once to a project, a loss and a lawsuit. After those relationships are created, any agent assigned a related task can see that the information might be relevant. Mapping relationships at entry therefore reduces repeated discovery work for later queries. The graph becomes reusable business context rather than a one-time migration artifact.
What graph-based modernization means for business
For companies running legacy estates, the practical effect is a different modernization sequence. Discovery of dependencies and business logic, usually a slow manual phase, moves into an automated mapping stage before rebuilding starts. That can shorten assessment timelines and make fixed-scope projects more predictable. Small firms with little documentation of old Delphi or Oracle Forms systems gain a way to recover hidden rules, while large firms can standardize context across many applications.
The approach also changes what buyers should verify before committing. A graph is only as complete as the code, data and rules ingested, so coverage of integrations and undocumented behavior matters. BotMan links unstructured information through entities such as a person, a project or a product, connecting messages and documents to known context. Buyers need to check how entities are defined, how errors in links are corrected, and what human review remains for regulated processes.
The marker to watch is whether BotMan demonstrates full autonomous migrations from ingestion to a working new architecture on production systems. The interview was part of coverage sponsored by Neo4j, with no editorial control over content, according to the disclosure. Repeated cases across Delphi, Visual Basic and Oracle Forms estates would show that graph context scales beyond pilots. Absence of such cases would leave the method as a promising discovery aid.
