• Written by Muhammed Abdulla NC, CEO

How long before AI shows results in research and development is a question I have heard many times in meetings, and it is hard to answer because the timeline for AI to show results can vary greatly depending on the complexity of the project and the data available. In my experience, the first fortnight of an AI project in research and development is often spent on connecting systems and preparing data, which is a crucial step but not the most exciting one for stakeholders who are eager to see results.

Setting the foundation

The first fortnight of an AI project is typically spent on setting up the infrastructure, connecting systems, and preparing the data. This includes connecting AsscherAi to the organisation's data sources, setting up user roles and access, and ensuring that the data is in a format that can be used by the AI system. While this may seem like a mundane task, it is essential to get it right in order to ensure that the AI system has access to the data it needs to provide accurate and useful insights.

This initial setup period is also a good opportunity to identify any potential issues or roadblocks that may impact the project timeline. For example, if the data is not in a suitable format, it may need to be cleaned or transformed before it can be used by the AI system. By identifying these issues early on, the project team can take steps to address them and ensure that the project stays on track.

The first question worth answering

The first question worth answering with AI in research and development is often one that has a clear and measurable impact on the organisation. For example, a shift supervisor may want to know which machines are most likely to require maintenance in the next week, or a category manager may want to know which suppliers are most likely to meet their delivery targets. By answering these types of questions, the AI system can provide valuable insights that can help the organisation to make better decisions and improve its operations.

In order to identify the first question worth answering, it is essential to work closely with stakeholders to understand their needs and priorities. This may involve conducting interviews or workshops to identify the key challenges and opportunities facing the organisation, and then using this information to identify the questions that are most worth answering.

Measurable change

A leading measure should move within a few months of implementing an AI system in research and development. For example, if the AI system is being used to predict machine failures, a leading measure may be the number of unexpected failures per month. If the AI system is being used to optimise supply chains, a leading measure may be the number of days it takes for goods to be delivered. By tracking these types of measures, the organisation can see whether the AI system is having a positive impact and make adjustments as needed.

It is also important to ensure that the leading measures are aligned with the organisation's overall goals and objectives. For example, if the organisation is trying to reduce costs, the leading measures should be focused on cost-related metrics such as reduction in maintenance costs or improvement in supply chain efficiency. By aligning the leading measures with the organisation's goals, the AI system can be used to drive meaningful change and improvement.

Common pitfalls

One thing that can make an AI project in research and development take twice as long is poor data quality. If the data is inaccurate, incomplete, or inconsistent, it can be difficult or impossible for the AI system to provide accurate and useful insights. For example, if the data on machine failures is not accurate, the AI system may not be able to predict failures correctly, which can lead to unnecessary maintenance or unexpected downtime.

Another common pitfall is not having a clear understanding of what the organisation is trying to achieve with the AI system. If the goals and objectives are not clearly defined, it can be difficult to identify the questions that are most worth answering, and the project may not have a clear direction. By avoiding these types of pitfalls, organisations can ensure that their AI projects in research and development are successful and provide meaningful benefits.

Limitations of AI in research and development

While AI can be a powerful tool in research and development, it is not a silver bullet and there are limitations to its use. For example, AI is not a replacement for human judgement and expertise, and it should be used to augment and support human decision-making rather than replace it. Additionally, AI is not suitable for all types of research and development projects, and it may not be the best choice for projects that require a high degree of creativity or innovation. For more information on how to get started with AI in research and development, visit our research and development page, or contact us to speak with one of our experts.