Intellectual property management in early-stage start-ups : patent searching for digital twin technologies

Abstract

Early-stage start-ups developing digital twin (DT) technologies face a demanding intellectual property (IP) management task: identifying relevant patents while relying solely on the search capabilities of free patent databases. DT inventions span multiple technical domains and are distributed across heterogeneous CPC classifications, while terminology in the field remains inconsistent. Limited in-house IP expertise and the basic search functions of free tools further reduce the reliability of keyword-based searches. Although hybrid and AI-assisted retrieval methods exist, they usually depend on commercial platforms unavailable to most start-ups.

This thesis develops a theory-ingrained analytical patent search framework tailored to early-stage DT ventures and implemented entirely with free patent information sources. The framework integrates CPC domain mapping, context-specific keyword strategies, and staged manual screening. It is demonstrated through a qualitative case study of Berrytron Oy, a Finland-based start-up developing Reality-Driven Simulation (RealSim), with expert input informing the interpretation of the results.

The findings show that a structured hybrid strategy can identify patents relevant to the functional architecture of RealSim even when terminology is inconsistent and DT-related inventions are dispersed across multiple CPC fields. The framework strengthens early-stage patent intelligence by clarifying technological proximity, identifying enabling technologies, and highlighting opportunities for differentiation, while remaining distinct from formal legal or freedom-to-operate assessment. Overall, the thesis provides a transparent and practically applicable method for conducting early patent searches in emerging and technically complex domains, particularly for early-stage start-ups operating under resource constraints.