• Written by Azhar Aliyar, Director

A research engineer checks the test results for a new material against the design specifications, but the data is scattered across different systems, making it difficult to get a complete picture. This is a common problem in research and development, where data is often siloed in different systems, making it hard to integrate and analyze.

Read paths, write paths and why the second needs more care

When integrating AI systems with research and development data, it's essential to consider the read paths and write paths. Read paths refer to how the AI system accesses the data, while write paths refer to how the AI system updates the data. While read paths are relatively straightforward, write paths require more care, as they can potentially create inconsistencies and errors in the data.

The reason write paths need more care is that they can create a second source of truth, which can lead to data inconsistencies and errors. For example, if the AI system updates the test results in its own database, but not in the original system of record, this can create a discrepancy between the two systems. To avoid this, it's crucial to ensure that the AI system updates the data in the original system of record, rather than creating a separate database.

Reconciling definitions that differ between systems

Another challenge in integrating AI systems with research and development data is reconciling definitions that differ between systems. For example, one system may define a "test result" as a specific set of data, while another system may define it differently. To reconcile these differences, it's necessary to create a common data model that defines the terms and concepts used across all systems.

This requires a thorough understanding of the data and the systems used in research and development. It's not just a matter of creating a common vocabulary, but also of understanding the context and nuances of each system. For instance, a term like "failure mode" may have different meanings in different systems, and it's essential to understand these differences to create a common data model.

The research and development data this depends on

The integration of AI systems with research and development data depends on a range of data sources, including test results and failure analyses, design review records, timesheets and allocation, and supplier assessments. Test results and failure analyses provide critical information about the performance and reliability of materials and systems, while design review records provide insight into the design process and decisions made.

Timesheets and allocation data provide information about the resources used in research and development, while supplier assessments provide insight into the quality and reliability of external suppliers. By integrating these data sources, AI systems can provide a comprehensive view of research and development activities and help identify areas for improvement. For more information on research and development, visit our research and development page.

What it changes about: failures rediscovered across programmes and effort invisible against project stage

The integration of AI systems with research and development data can change the way failures are identified and addressed across different programs. By analyzing data from multiple sources, AI systems can identify patterns and trends that may not be apparent from a single data source. This can help researchers and developers identify potential failures earlier, reducing the risk of costly rework or project delays.

Additionally, the integration of AI systems with research and development data can make effort more visible against project stage. By tracking resources used and progress made, AI systems can provide a clear picture of how projects are progressing and where resources are being used. This can help researchers and developers optimize their workflows and allocate resources more effectively.

Measured with: failures found before validation, effort against stage, estimate revision count

The effectiveness of AI systems integration with research and development data can be measured in several ways, including the number of failures found before validation, effort against stage, and estimate revision count. By tracking these metrics, researchers and developers can evaluate the impact of AI systems on their workflows and identify areas for further improvement.

For example, if the number of failures found before validation increases, this may indicate that the AI system is effectively identifying potential issues earlier in the development process. Similarly, if effort against stage decreases, this may indicate that the AI system is helping researchers and developers optimize their workflows and allocate resources more effectively. If you have any questions about AI systems integration, please contact us for more information.

What this does not do

The integration of AI systems with research and development data is not a replacement for human judgment and expertise. While AI systems can provide valuable insights and analysis, they are not a substitute for the expertise and experience of researchers and developers. Additionally, AI systems integration is not a one-time process, but rather an ongoing effort that requires continuous monitoring and maintenance to ensure that the data remains accurate and up-to-date.

Furthermore, AI systems integration is not suitable for all research and development activities. For example, activities that require a high degree of creativity or human intuition may not be well-suited for AI systems integration. In these cases, human judgment and expertise are essential, and AI systems may not be able to provide the same level of insight or analysis.