• Written by Muhammed Abdulla NC, CEO

A research and development manager stares at a spreadsheet, trying to understand where the budget for their AI project has gone, and why the costs seem to be accumulating in unexpected places. The initial estimate for the AI cost in research and development had seemed reasonable, but now the numbers are not adding up.

Integration and data work as the dominant line

The largest portion of the budget is often spent on integration and data work, rather than the AI model itself. This is because the AI model requires a large amount of high-quality data to function effectively, and integrating this data into the existing systems can be a complex and time-consuming process. Many organisations underestimate the amount of work required to prepare their data for AI, and this can lead to cost overruns.

A typical example of this can be seen in the case of a product development team, where a data scientist spends a significant amount of time cleaning and preprocessing the data before it can be used to train the AI model. This work is essential to the success of the project, but it is often not accounted for in the initial budget.

Run cost against query volume

Another area where costs can accumulate is in the run cost of the AI system, which is often directly related to the volume of queries it receives. As the number of queries increases, so does the cost of running the system, and this can quickly add up. Most people get wrong the idea that the cost of AI is a one-time payment, when in reality it is an ongoing expense that needs to be carefully managed.

For instance, a shift supervisor checks the maintenance log and notices that the AI system is receiving a large number of queries from the production team, which is increasing the run cost of the system. To mitigate this, the supervisor can work with the production team to optimise their queries and reduce the volume of requests.

The cost of a wrong answer nobody caught

The cost of a wrong answer from the AI system can be significant, especially if it is not caught and corrected in time. This can lead to a range of negative consequences, from wasted resources to damage to the organisation's reputation. It is essential to have a robust testing and validation process in place to ensure that the AI system is providing accurate and reliable answers.

A wrong answer can have serious consequences, and it is essential to consider this when evaluating the AI cost in research and development. The cost of correcting a mistake can be much higher than the cost of preventing it in the first place, and this should be factored into the budget.

Attributing spend to a process rather than a model

When evaluating the AI cost in research and development, it is essential to attribute the spend to the process rather than the model. This means considering the entire workflow and how the AI system fits into it, rather than just focusing on the cost of the model itself. By taking a more holistic approach, organisations can get a better understanding of where their budget is being spent and make more informed decisions.

For example, a category manager compares two suppliers and notices that one of them is using an AI system to optimise their logistics. The manager can use research and development techniques to evaluate the effectiveness of this system and determine whether it would be a good investment for their own organisation.

What this does not do

The AI cost in research and development is not a one-size-fits-all solution, and there are certain situations where it may not be the best choice. For instance, if an organisation is looking to implement a simple automation solution, the cost of an AI system may be prohibitively expensive. In such cases, it may be more effective to use a different approach, such as a rule-based system.

Organisations should carefully evaluate their needs and consider whether an AI system is the best solution for their particular use case. If you are considering implementing an AI system and want to learn more about the potential costs and benefits, you can contact us to discuss your options in more detail. The key to successful AI implementation is to carefully consider the costs and benefits and to make an informed decision based on your organisation's specific needs.

Ultimately, the AI cost in research and development is just one part of the overall equation, and organisations must consider a range of factors when evaluating the potential benefits and drawbacks of an AI system. By taking a careful and considered approach, organisations can make the most of their investment in AI and achieve their research and development goals.