Neuron Data

Neuron Data launched in 1985 to solve a critical bottleneck: expert knowledge existed only in specialists' heads, inaccessible to organizations that needed it. ​​‌‌‌‌‌‌‌​‌‌​​‌​​​​​​‌‌​‌‌‌​​​‌‌Manufacturing plants, financial institutions, and diagnostic centers experienced this acutely—they couldn't scale expertise or ensure consistency without the original expert present. The problem was measurable: companies tracked costly delays, inconsistent decisions, and expertise loss when specialists retired. Alternatives existed but were crude: decision trees on paper, custom programming, or hiring more experts. Neuron Data's Nexpert shell promised to capture expert logic in software, making it replicable and portable. However, the company missed critical warning signs. Building expert systems proved far harder than anticipated—extracting and encoding human expertise required months of specialist interviews, and systems became brittle when encountering novel situations. The market also shifted unexpectedly toward general-purpose databases and client-server architectures rather than specialized shells. Neuron Data pursued aggressive platform expansion (Macintosh, Windows, Unix, VAX) without establishing product-market fit, diluting resources across incompatible systems. They underestimated how quickly the underlying technology would commoditize.

Neuron Data launched Nexpert in 1985 as a stripped-down expert system shell for Macintosh, deliberately omitting advanced features competitors offered to ship quickly. The MVP focused on backward chaining logic with a clean interface, reaching market within months of founding. This speed proved both asset and liability. While rapid porting to PC and Windows positioned them early in emerging platforms, the company prioritized breadth over depth—spreading resources across Mac, PC, VAX, and Unix variants before establishing product-market fit on any single platform. The warning signs emerged gradually. By pursuing aggressive cross-platform expansion, Neuron Data diluted engineering focus and customer support. Each port introduced compatibility issues and maintenance burdens. They competed against entrenched players like Exsys and CLIPS without a clear differentiation strategy. The deliberate minimalism that enabled speed became a liability as customers demanded richer functionality. Neuron Data's execution velocity masked fundamental questions about whether expert systems would sustain commercial viability, ultimately leading to their acquisition by Dun & Bradstreet Software in 1989.