Artificial intelligence: a manufacturer’s best assistant?

It’s changing the way manufacturers work too, enabling smarter decision-making and automation. Large-scale manufacturers are already using it for tasks like predictive maintenance, quality control and simulating prototypes via digital twins. Rolls Royce, for instance, is embedding AI into its engine digital twins, so it can spot patterns and potential problems before they occur. 

Around three quarters of manufacturing firms are planning to increase AI investment in 2025 but adoption is slow with only a third of companies actually using it. Some of the most common barriers include skills gaps, complexity, costs and technology challenges. 

If you're a small-to-medium-sized enterprise (SME), you don’t need to start big – AI can be a great assistant when embedded into your day-to-day systems such as enterprise resource planning (ERP) software. 

Key summary 

• AI is increasingly embedded into manufacturing processes to save time, improve efficiency and deliver innovations
• Manufacturers need the right processes and tools in place before investing in AI like an ERP and high-quality data
• Experts at ECI are developing innovative AI solutions, embedded into ERPs, to make the technology more accessible and less complex for SME manufacturers

The UK government is ramping up AI adoption across the UK with its AI Opportunities Action Plan. It includes 50 recommendations to improve the UK’s AI sector and encourage businesses, including those in manufacturing, to adopt it. 

To use AI as an assistant, it needs to understand your business and integrate with your tools. And for that you need quality data from across your business – everything from production to financials to sales. The more quality data you give to AI, the more it learns and improves.

AI offers opportunities we can’t let slip through our fingers, and these steps put us on the strongest possible footing to ensure AI delivers in all corners of the country, from building skills and talent to revolutionising our infrastructure and compute power.

Matt Clifford CBE, Prime Minister's AI Opportunities Advisor 
 

Your ERP is the starting point for AI implementation because it contains data you capture from the business. The data within it provides a reliable ‘single source of truth’ which AI applications can use to make better decisions for your business. 

Here are some ways you could use AI as an assistant: 

Manage inventory levels: Smart shelves with embedded weight sensors send data to your ERP, so the AI tool can compare it against required stock levels. You can then set a trigger for it to be automatically ordered or send an alert to the purchasing department to reorder. You’ll also get insights such as fast moving inventory or products for better stock management and demand forecasting. 

Predictive maintenance: Pressure or heat sensors installed on or near equipment will help you predict any problems. For example, if a machine is overheating, AI can detect it early on by analysing the data and sending your team an alert. It helps you predict breakdowns and schedule maintenance. 

Sustainability: AI can monitor energy and water usage, and waste levels, to spot any inefficiencies. Manufacturers can then take practical steps to cut down, for example, switching machines off if not in use. 

AI doesn’t eliminate manufacturing jobs – it’s more of an assistant for manufacturers who can use it to speed up processes like data analysis. It still needs complete supervision by a human because AI can make mistakes if data is incomplete/missing. It does give manufacturers more time to work on value-added tasks and areas like creativity, which AI cannot replicate. 

The cost and time-saving benefits AI can offer make it an exciting prospect. But it’s important to remember that it won’t automatically solve your challenges if you don’t have the right data or prompts. 

To get the most from it, and avoid misleading results, you first need an ERP that centralises and gives you oversight of your data. 

What to consider:

1. AI relies on high-quality, current data to work. If the data is missing or incomplete, it can lead to unreliable results.
2. An ERP is essential to adopting AI because it provides a central platform for your data, helping to prevent silos and keeping it up to date – some ERPs also have AI capabilities.
3. AI is prone to hallucinations and errors if it doesn’t have the right information. If an application automatically orders materials without human oversight, you could end up with too much expensive stock sitting in your warehouse.
4. Without the right safeguards, you may not be able to explain the rationale behind AI decisions (the ‘black box’ problem). This could impact the accuracy of your forecasts, leading to either over or understocking.