How long before AI shows results in quality?
Implementing AI to improve quality involves setting up the system, identifying key questions, and tracking leading measures, such as defect rates or customer satisfaction, to drive tangible outcomes and measurable change, as seen when a category manager uses AI to compare supplier quality.
A dashboard that nobody opens is a common sight in many organisations, a reminder that the introduction of AI is not a guarantee of immediate results in quality. The business consequence of a poorly planned AI timeline is a lack of measurable change, which can lead to disappointment and skepticism about the value of AI. In my experience, the mechanism that drives this lack of progress is the failure to establish a clear sequence of events, from connection to tangible outcomes.
Initial Setup
The first fortnight of AI implementation is often spent on technical setup, such as connecting data sources and configuring the system. This is a necessary step, but it is not a guarantee of success. I have seen many organisations get bogged down in this phase, trying to perfect the setup before moving on to the next step. However, this approach is wrong, as it prioritises technical perfection over business outcomes. A better approach is to focus on getting the system up and running, and then iterating on the setup as needed.
A shift supervisor, for example, may be tasked with checking the maintenance log to identify areas where AI can be used to improve quality. This initial step is crucial in establishing a clear understanding of the organisation's data and identifying potential areas for improvement.
The First Question
The first question worth answering is often one that has a clear and direct impact on quality. For instance, a category manager may want to know which suppliers are consistently delivering high-quality products. Answering this question can help the organisation to identify areas where it can improve its quality control processes. Most people get wrong the idea that AI should be used to answer complex and abstract questions, but this approach is misguided, as it can lead to vague and unactionable insights.
In contrast, focusing on concrete and specific questions can help to drive tangible outcomes and measurable change. By using AI to answer questions like this, organisations can start to see real improvements in quality, which can have a direct impact on the bottom line.
Leading Measures
A leading measure should move within a few weeks of implementing AI, as the organisation starts to see the impact of its efforts. For example, a reduction in defect rates or an improvement in customer satisfaction scores can be an indicator that the AI is having a positive effect. However, this is not always the case, and some organisations may take longer to see results. The key is to establish clear metrics and track progress regularly, using tools like production monitoring systems to stay on top of performance.
When a leading measure moves, it is a sign that the organisation is on the right track, and that the AI is starting to have a real impact on quality. This can be a powerful motivator, as it shows that the efforts of the team are paying off and that the organisation is making progress towards its goals.
Delays and Challenges
What makes it take twice as long to see results from AI is often the failure to address underlying issues, such as data quality problems or inadequate training. These challenges can be significant, and can delay the implementation of AI by weeks or even months. Additionally, organisational barriers, such as resistance to change or lack of buy-in from stakeholders, can also slow down the process. To overcome these challenges, organisations need to be proactive and address these issues head-on, which can be a difficult and time-consuming process.
For instance, an organisation may need to invest in data cleaning and preparation, or provide training to employees on how to use the AI system effectively. By addressing these challenges, organisations can help to ensure that the AI is implemented successfully and that results are seen as quickly as possible. If you are experiencing delays or challenges with your AI implementation, you can contact us for support and guidance.
Limitations and Exceptions
This approach to AI implementation does not work for every organisation, and there are some cases where it may not be the best choice. For example, organisations with very complex or bespoke systems may need a more tailored approach, and may require additional support and guidance. In these cases, a more traditional implementation approach may be necessary, which can be time-consuming and resource-intensive.
Additionally, some organisations may not have the necessary infrastructure or resources to support AI implementation, which can make it difficult to see results. In these cases, it may be necessary to invest in additional resources, such as hardware or personnel, before implementing AI. By understanding the limitations and exceptions of this approach, organisations can make informed decisions about how to proceed with their AI implementation.
Frequently asked questions
What is the first step in implementing AI to improve quality?
The first step is to set up the system, connecting data sources and configuring it to get up and running.
How do I identify the right questions to answer with AI?
Focus on concrete and specific questions that have a clear impact on quality, such as identifying suppliers with consistently high-quality products.
What are leading measures and how do I track them?
Leading measures are indicators of progress, such as defect rates or customer satisfaction, and can be tracked using tools like production monitoring systems.
What are common challenges that can delay AI implementation?
Common challenges include data quality problems, inadequate training, and organisational barriers, which can be addressed by investing in data cleaning and preparation, and providing training to employees.