Business

Untapped opportunities of AI in moving towards smart manufacturing

Smart manufacturing requires modernized sales and operations planning processes through artificial intelligence (AI) -based digital solutions.
Smart manufacturing requires modernized sales and operations planning processes through artificial intelligence (AI) -based digital solutions.

Manufacturing firms are investing heavily in new technologies to capitalize on the opportunities of smart manufacturing. While many firms focus their investments on further optimizing their production processes, front-running firms have understood that smart manufacturing requires them also to modernize their sales and operations planning processes through artificial intelligence (AI) -based digital solutions.  

Amidst competitive pressure from low-cost rivals having factories in Asia and increasingly in Africa too, manufacturing firms have begun to invest in new technologies that enable the firms to boost efficiency, adaptability, and flexibility of their production processes – technologies often grouped under the broad term of smart manufacturing. BMW, for example, has systematically invested in technologies for robotics applications in its different production processes.

In the C-suite, such technology investments are seen to be critical for maintaining the competitiveness of firms. After all, the ability to produce high-quality premium products for niche markets has been the recipe of success for many European manufacturing firms, such as the Finnish Vaisala, a manufacturer of environmental and industrial measurement instruments (we’ve been happy to work with them for years!).

While manufacturers have made good progress so far by bringing new technology to production, a seasoned strategist is likely to ask if the investments made so far will be sufficient for maintaining competitiveness in the long term. From this perspective, a common risk we see arises from the fact that many manufacturing firms have so far invested in technology geared towards production optimization – including both integrated systems (e.g. MES) and focused solutions (e.g. machine vision and machine learning-based quality assurance solutions).

Although production optimization has a long legacy in manufacturing and will continue to be important for manufacturing firms also in the future, the risk is that production optimization alone will not be sufficient for long-term competitiveness in the age of smart manufacturing. 

For being able to reap the commercial benefits of earlier investments in smart manufacturing, it is critical that manufacturing firms are able to also modernize functions and task domains that steer core production processes and set demands for them. Sales and operations planning, in particular, is a critical business process that not only has a direct impact on firm performance, but that also often becomes a bottleneck when firms try to benefit from more adaptable and flexible manufacturing systems. When firms adopt smart manufacturing technologies, they also need to increase the pace and adaptability of their sales and operations planning processes.

Yet, a challenge related to this is that people responsible for these processes (e.g. production planners) still commonly lack modern artificial intelligence (AI) -based tools that enable them to run the planning processes faster and in a more flexible manner. This is a critical condition for capitalizing on the investments made in smart manufacturing that has received little attention so far.

Next, I’ll elaborate on how firms moving towards smart manufacturing can leverage AI-based digital solutions to modernize sales and operations planning processes.

It’s good to be explicit that we by no means believe that tools alone would bring the aspired business impact. Rather, we believe that the value of new types of tools can be realized if these tools enable the different people responsible for the sales and operations planning process to come together, to form more forward-looking and more strategic plans, and also to reflect the way they do their work and run the planning processes, more generally speaking.

New requirements for manufacturing firms’ sales and operations planning 

Shift towards smart manufacturing will not change the established managerial wisdom that key to driving long-term firm performance is having a portfolio of products that are profitable (over their product life cycle) and have demand in selected markets.

However, what is expected to change as firms move towards smart manufacturing is that markets become more fragmented and also more volatile. When manufacturers adopt smart manufacturing technologies to increase the adaptability and flexibility of their production processes, they are able to produce more tailored and smaller batches of products – a key capability for targeting tailored products to more granular customer segments or to short-lived market trends.

As such, the long-term opportunity associated with smart manufacturing arises from a firm’s increased ability to align its product portfolio in a more continuous and agile manner vis-à-vis more granular and volatile market needs – at a premium price. 

More continuous and agile product portfolio management, however, represents a significant challenge for manufacturing firms.

Although manufacturing firms have traditionally been exceptionally good in driving production efficiency through the use of sophisticated data and information systems, the processes that steer production – product management as well as sales and operations planning processes in particular – have so far utilized relatively less smart technology. For example, many manufacturing firms still run their sales and operations planning process through a series of planning meetings and spreadsheet calculations of operations planning engineers – technologies dating back to the 1980s.

As smart manufacturing technologies become more common in the 2020s, this will gradually shorten product life cycles and increase demand volatility and, as a consequence, manufacturing firms will very likely need to invest in modern data and AI tools that enable them to shorten their sales and operations planning frequency and horizon – thereby paving the way for more dynamic business and its different elements such as dynamic pricing, demand prediction, and automatized lead generation. 

Over the last years, Reaktor has been collaborating with digital frontrunners in different industries to build modern digital solutions that leverage different types of data (e.g. external customer data, market data) and machine learning (ML) algorithms. Such digital solutions are increasingly being used to steer production and logistics operations in a more automatic manner, for example in retail. While they are not yet commonplace in manufacturing, we believe that similar solutions could be utilized to speed up and improve the planning and management processes of manufacturing firms. 

Modernize your sales and operations planning through focused digital solutions incorporating machine learning algorithms in their core

To understand the business opportunity associated with modernizing sales and operations planning, it is useful to first look at a core element of a modern sales and operations planning process: machine learning. As you probably know, machine learning is a branch of artificial intelligence where a software system can learn from data, identify patterns and make predictions or simple decisions with minimal human intervention.

A machine-learning algorithm could, for example, do the same magic that a production planning engineer would do in her or his spreadsheet – with the exception that a machine learning algorithm can crawl through much more massive amounts of data and identify complex patterns that a human brain typically simply can not detect.

By doing so, a machine learning algorithm can, for example, reveal nonoptimal assumptions underpinning the way sales and operations planning is currently being done – such as an assumption that certain batch sizes are most profitable while this is actually true only for certain products or factories – and also come up with new, more highly performing production plans that humans would not have been able to develop.

Machine learning algorithms are most useful in contexts where complexity is high.

Think of a manufacturing firm operating in tens of markets, having a product portfolio comprising thousands of products, and having a network of factories. In such a context, optimizing firm performance requires a solution that is able to estimate what combination of production plans (of different factories) best match the products that will be produced to different needs in different markets at a price that maximizes the KPIs of the firm (e.g. profitability).

The key benefit of machine learning algorithms is that they can deal with such complexity and make recommendations about what products each factory should produce. Furthermore, machine learning algorithms also learn from new data and they, as such, continuously improve firm performance as products get produced and the firm collects data about revenue from different markets in which they operate. Machine learning algorithms can thus do complex calculations, predictions, and recommendations and deliver recommendations on demand – even thousands of recommendations every second. 

The modern way for deploying such machine learning algorithms to production processes is to package them as production-grade focused digital solutions that do some particular tasks well and fast. These digital solutions do all the hard calculations, predictions, and recommendations in the background and deliver key results – e.g. recommendations about product portfolio – to the user through a user interface.

In a manufacturing context, machine learning algorithms as digital solutions would enable a production planning engineer, for example, to search for the optimal production plans for each factory in a faster and more effective manner by means of “outsourcing” most of the number-crunching work she has done in the past in her spreadsheets to the machine learning-based digital solutions. By doing so, the production planning engineer could focus her time, effort, and expertise on making the final decisions based on recommendations and insights brought by the machine learning algorithms. 

Enable your planning professionals to do their work better by providing them with modern AI tools

As manufacturing firms invest in smart manufacturing technologies to boost the efficiency, adaptability, and flexibility of their production processes, it becomes critical to look at how they also further develop the functions and task domains that steer core production processes and set demands for them.

In this blog text, we have shed light on how digital solutions built on top of machine learning algorithms could enable professionals running sales and operations planning processes to focus on the most important matters – making better decisions based on recommendations and insights brought by AI and thereby boosting firm performance. 


At Reaktor, we work with clients to help them understand the impact of digital innovations on their business and leverage business opportunities arising from new technologies such as machine learning. 

 

 

Let's build something together

Riku Ruotsalainen

Principal consultant, Strategy and Business Design

Markku Myllylahti

Business Director, Manufacturing

*Emails are in the format firstname.lastname@reaktor.com

Never miss a post