As the digital transformation of how we do business makes it increasingly difficult to create value using traditional approaches - machine learning programs offer a new way forward.

This definitive guide is designed to give you a simple understanding of how AI, and more specifically Machine Learning (ML), can give you a competitive advantage across the full range of operations of your manufacturing business – from promotion (leads, sales and marketing) to production and provision of service, and projections for business innovation and development.

We’ll explain what’s possible and give examples demonstrating business benefits. And, because ML isn’t a one-size-fits-all solution, we’ll follow it up with some practical tips and resources to help you decide where to apply ML applications in your business for the best ROI.

Technological advances have affected the entire business process – from how we engage with customers and their shifting demands, to implications for production economics, and expectations around speed and availability of supply. Add to that, the proliferation of smaller agile market entrants; and it has never been more critical for manufacturers to embrace technology to build a competitive advantage.

The reality is that those that can’t adapt to take advantage of the information, efficiencies and economies offered by AI applications will soon be left in the wake of those with a clear AI strategy.

If you’re hidebound, if you’re stuck to the old way and don’t have the capacity to digitalize manufacturing processes, your costs are probably going to rise, your products are going to be late to market, and your ability to provide distinctive value-add to customers will decline.

Stephen Ezell, Global Innovation Policy, Information Technology and Innovation Foundation

Discover the AI Business Advantage

Why is AI so crucial in today’s business environment?

AI has irrevocably changed our relationship with information – how we produce it, collect it and apply it. Every single instance of data that your organisation generates holds potential value – for cutting costs and creating business revenue.

And what’s more, according to the World Economic Forum 2017 white paper on technology and innovation, “70% of captured production data goes unused”. That’s an incredible untapped resource!

AI analytic strategies involving Machine Learning (ML) make sense of the stream of data generated by factories, operations and consumers; empowering manufacturers with the necessary insights to make decisions for maximum impact, and the flexibility to pivot to meet changes and challenges head on.

Every single instance of data that your organisation generates holds potential value – for cutting costs and creating business revenue.

ML knows how to optimise your business

Machine learning takes things a step further than the usual IT problem-solving approach – instead of algorithms, computers are fed massively huge data sets along with an instruction starting point from which to extrapolate.

Machine learning models can recognise patterns in a vast array of data and each new piece of information is used to update the existing knowledge base and optimise functionality. If the model gets it wrong, that new data allows it to adapt to do better the next time.

Machine learning allows us to unravel those patterns that would be difficult or impossible for people to identify.

Alp Kucukelbir, co-founder and chief data scientist, Fero Labs (Source)

By applying ML against the data coming in from the manufacturing process you gain the knowledge necessary to harness opportunities for refining your processes for success.

Using these insights ML programs can improve operational efficiencies, speed production, optimise equipment performance, minimise waste and reduce maintenance costs.

Read on for practical use cases covering the breadth of the manufacturing process.

  1. Boost Your Marketing and Sales with Proven AI Techniques
  2. Optimise Your Production and Maintenance Processes with ML Insights that Save You Time & Money
  3. Maximise Supply Chain Efficiencies & Enhance User Experience with Real-Time AI Feedback
  4. Innovate with AI-Driven Smarter R&D, Business Modelling & ForecastingI

These all contain examples that are already reaping benefits for forward-thinking business owners. You could be among them.

1

Boost Your Marketing and Sales with Proven AI Techniques

Marketing and sales has always been a data-driven field. With advances in AI that amount of data has increased exponentially. In fact (to quote The Royal Society in their paper Machine learning: the power and promise of computers that learn by example) “almost 90% of the world’s data is estimated to have been produced within the last five years”.

Since the success of an ML learning application is largely dependent on data set size, sales is an area where large benefits can be gained. It’s no surprise then that “84% of marketing organisations are implementing or expanding AI and machine learning in 2018”.

I think artificial intelligence uniquely gives us the ability to deliver on the promise of the customer at the centre. The machine can now deliver the right communication to that customer based on what you can infer about them.

Mike Handes, Customer Success Director, Marketo

For marketing and sales these gains are particularly relevant in relation to customer demand and value chain economics. Advanced analytics and ML can provide marked advantages in tailored customer experience and personalisation, as well as targeted customer engagement.

Here are some areas of marketing and sales that have ML models in operation today.

See who is using your products and what they are saying about your brand

Text extraction and summarisation of trending news to develop a picture of consumer opinion

E.g. Get a daily summary of what the world is saying about your organisation and a picture of general public sentiment

Robotic process automation to track social media triggers

E.g. Have AI optimise your marketing content based on consumer interaction

ML-powered computer vision for product recognition

E.g. Analyse images in a video stream to see how many people have your product

Recurrent neural networks (RNN) for text generation i.e. product names, campaigns

Statistics:

57% of the buying process is completed before a first interaction with sales. AI sales bots can hone in on customer ‘intend’ signals and have proactive answers for initial queries about pricing, product features, or contract terms. (Source)

ML models can hone your marketing content with clearer targeting

Bandit tests (instead of A/B tests) for immediate results in extracting and classifying relevant content

E.g. Replace traditional A/B split testing with an experimental machine learning model for less wastage during the testing phase

Automated propensity models for cross-sell and up-sell strategies

E.g. Know the right products to present at the right time to optimise purchase behaviour

Clustering for customer segmentation and discovery

Reinforcement learning for sequential marketing decisions

E.g. Know your optimised process for customer engagement. Know if an email after the first call works better than a second call or snail mail. Discover your best selling script

Ad optimisation

E.g. Put the right ad in front of the right audience at the right time

Statistics:

40% marketers will use AI to enhance content throughout the customer journey. (Source)

By 2020, real-time personalised advertising across digital platforms and optimised message targeting accuracy, context and precision will accelerate. (Source)

2,900 messages/day – consumers can only remember 4 (Source)

#1 source of customer disengagement: irrelevant content (Source)

80% marketers agree personalised content works better than generic (Source)

Optimise your customer experience offering with ML that anticipates and personalises your customer interactions both pre- and post-sales

ML risk prediction and intervention models to reduce customer churn

E.g. Know the actions you need to take and when to avoid losing a customer

Contextual content, offers and incentives to optimise the marketing mix

Voice-based digital assistants and dialog systems – customer experience automation

E.g. Reduce your reception overhead by creating a virtual assistant

Text classification to process customer feedback/ sentiment for user insight and personalisation

E.g. Recognise the tone of discussions in an email or chatbot and provide suggested responses

Voice-based search via text-to-speech (TTS) and speech-to-text (STT)

E.g. Replace dictation to capture your meetings in real time

Recommendation engines and targeting

E.g. Give the customer the right information at the appropriate time to improve customer satisfaction

Statistics:

75% of enterprises using AI and machine learning enhance customer satisfaction by more than 10% (Source)

57% of enterprise executives believe the most significant growth benefit of AI and machine learning will be improving customer experiences and support. (Source)

58% of enterprises are tackling the most challenging marketing problems with AI and machine learning first, prioritising personalised customer care, new product development. (Source)

1,000% leap in the use of virtual customer assistants by 2020. (Source)

90% businesses say customer experience is their competitive battleground. (Source)

ML models can ensure your numbers – whether it’s pricing or sales projections – stay on point

Regression models for dynamic pricing

E.g. Optimise your pricing so you can maximise your profit based on consumer interactions

Predictive models for sales leads and ideal customer profiles

E.g. Stop taking your reports on future sales from over-optimistic sales reps and obtain statistical predictions based on current sales and pipeline activities

RFM (recency, frequency and monetary) modeling to define sales projections of specific customer segments and microsegments

E.g. Understand what factors – weather, time of year, advertising spend by competitors – influence your sales forecast, and adjust production accordingly

Statistics:

Improving demand forecasting, assortment efficiency and pricing in retail marketing have the potential to deliver a 2% improvement in Earnings Before Interest & Taxes (EBIT), 20% stock reduction and 2 million fewer product returns a year. (Source)

Cut down on repetitive admin tasks and boost output with ML shortcuts

Robotic process automation for reading emails

E.g. Have a model sort your emails and prioritise your focus

Automatic data entry/visualisation for templated reports

ML quota setting models based on patterns rather than one-size-fits-all

E.g. Learn when to expand and contract production based on market signals

AI bots to identify at risk accounts and untapped connections

Automated scheduling

Opportunity scoring to improve win rates

2

Optimise Your Production and Maintenance Processes with ML Insights that Save You Time & Money

Manufacturers statistics show that fifteen billion machines are currently connected to the internet, and in 2020, the number will pass 50 billion. But, according to Intel, “while many manufacturers have sensors in place, few are using them to glean revenue-driving insights”.

ML benefits for the factory floor are dependent on manufacturing execution systems (MES) – these allow for communication between machines and assets by connecting them via a common infrastructure.

These real-time insights are what machine learning algorithms can process in order to boost industrial output, prevent costly machine breakdowns, and reduce waste. This is not only good for your bottom line, it makes for a more environmentally-friendly and sustainable production process.

Accenture predicts that as AI-powered machines eliminate faulty machines and idle equipment, manufacturers will experience consistently rising rates of return, resulting in equally dramatic profit increases of 39 percent by 2035

(Source)

There are a lot of gains to be made using ML to streamline production processes, as demonstrated in the following.

Use ML to get a precise picture of your production process

Monitor environmental conditions at plant level and implement adjustments

E.g. Discover how temperature changes affect your production quality and outputs

Continuous transmission of operational feedback in real time

Real time monitoring of shop floor operations

Reduction in unplanned downtime

E.g. Have a model optimise your maintenance program to reduce downtime

Optimising build production line

E.g. Analyse in intricate detail all the steps and time taken on your production line floor and provide recommendations for improvement

Reduce resting costs

Asset management – improve asset tracking accuracy

E.g. Add image recognition to your identification and storage process to improve accuracy

Process visualisation and automation

E.g. Have every item QA’d, rather than 1 in 20 to reduce defects and reduce returns

Insights into production schedule performance and machine-level loads

Real time linking of scheduling with shop floor operations

E.g. Only build what is predicted to be ordered

ML can identify where a small tweak can offer huge results for quality improvement

Reducing warranties

ML to detect and predict manufacturing failures

Computer vision to find microscopic defects

Improved quality management

Root-cause analysis

Process and quality optimisation

Integrating OEE (overall equipment effectiveness) by the asset to improve yield rates

Automated issue identification

Statistics:

Automating quality testing using machine learning is increasing defect detection rates up to 90% (Source)

ML allows you a proactive approach to maintenance that saves time and money

Predictive and preventative maintenance to prevent equipment failure with real-time preemptive solutions

Reduce test and calibration time

Statistics:

“Predictive analytics have delivered Kaesar a 60% reduction in unscheduled equipment downtime as well as an estimated annual savings of $10 million in break-fix costs, as the company can better predict its inventory needs.” (Source)

“If it takes a 3 minute phone call to report an issue, you’ve lost $70,000 just telling someone you have a problem.” (Source)

“McKinsey found that AI-enhanced predictive maintenance of industrial equipment will generate a 10% reduction in annual maintenance costs, up to a 20% downtime reduction and 25% reduction in inspection costs”. (Source)

“A manufacturer was able to achieve a 35% reduction in test and calibration time via accurate prediction of calibration and test results using machine learning”. (Source)

You can use ML to identify opportunities for cost-saving waste reduction and in the process create a win for the environment

Reduce scrap rates

Reduce energy consumption – environmental gains

Reduction in raw materials – increased sustainability

Statistics:

“Fero Labs is able to predict mill scaling with an accuracy of 78-100 percent, according to Kucukelbir, reducing it by 15 percent.5”. (Source)

“When setting up their AI platform, Fero Labs increased sensor data usage 40 times over simply by feeding previously unused information to their algorithms—offering comprehensive insights into factory activity without installing any new equipment.” (Source)

3

Maximise Supply Chain Efficiencies & Enhance User Experience with Real-Time AI Feedback

According to Price Waterhouse Cooper’s 2020 analysis of digital factories, three out of four respondents set up digital factories to react to customer preferences more quickly. This is perhaps not surprising given that customer focus is now driving production decisions more strongly than labour costs. When it comes to inventory and supply chain – real-time logistic strategies are key to being able to give customers the responsiveness and customisation they have come to expect.

Here are some of the reasons that smaller operators can be market competitive with ML.

Use ML to optimise everything from inventory distribution to delivery routes

Automated inventory optimisation to scale across distribution locations and factor variables affecting demand and time-to-customer delivery performance

Dynamic inventory replenishment

Reduction in supply chain forecasting errors

Real-time visibility into every machine making every component across supply chains

Cut transport and warehousing costs

Savings in supply chain admin

Demand forecast accuracy

ML algorithms for procurement, strategic sourcing and cost management

Demand forecast accuracy to reduce energy costs and negative price variances

Analysis of resource utilisation

Optimised delivery routing

E.g. Have a model create the delivery route to optimise on fuel and time

Improve your customer responsiveness with ML insights

Improving product availability

Faster intelligence processing for greater responsiveness to changes in consumer demand

Synchronisation to speed time-to-market

Delivery of customised product offerings

ML for price elasticity

Improving product availability

Faster intelligence processing for greater responsiveness to changes in consumer demand

Synchronisation to speed time-to-market

Delivery of customised product offerings

ML for price elasticity

Statistics:

McKinsey predicts machine learning will reduce supply chain forecasting errors by 50% and reduce lost sales by 65% with better product availability. (Source)

Automating inventory optimisation using machine learning has improved service levels by 16% while simultaneously increasing inventory turns by 25%. (Source)

4

Innovate with AI-Driven Smarter R&D, Business Modelling & Forecasting

New technology has created a wealth of opportunities for innovation – new markets have opened up along with a variety of new methods for consumer engagement. The power of machine learning is in its ability to identify patterns, apply solutions and learn from the results. This big picture view of your operations is ripe with opportunity for identifying and testing new directions and innovative approaches. And with generative design and virtual models, experimentation is risk-free.

There are so many ways you can use ML to test your ideas before taking the leap.

Enjoy risk-free innovation using ML to gain new perspective on your business potential

Statistics:

Improving the accuracy of detecting costs of performance degradation across multiple manufacturing scenarios reduces costs by 50% or more. (Source)

Identify new business models e.g. shifting costs from capital to operating expenses

Data-driven decision making for better operational decision-making

Determine cost behaviours across multiple scenarios

Generative design to test and learn from each iteration

Digital twins (of factory/ product/ production asset) – remote virtual model to learn and uncover opportunities

Knowledge capture for integrated planning

Determining cost behaviours across multiple manufacturing scenarios

So, What’s Your AI Strategy?

A lot has been written about the implications of AI for manufacturing. Naturally these documents tend to focus heavily on factory floor processes, but it’s important to remember that not all AI benefits relate directly to the production line. We hope this article has opened your mind on the “whole of business” opportunity AI poses for your organisation.

The starting point is a clear AI strategy, with an eye for leveraging machine learning applications that can best benefit your business. So ask yourself what data are your business operations generating and where can that be directed for short-term and long-term gains? And then, talk to technology experts with machine learning expertise – it’s not too late to find your digital competitive advantage.

Keen to Find Out More?

Check out these useful articles.

ML for marketing:

ML for production, maintenance and supply chain optimisation:

ML for business innovation:

A tailored solution that drives real growth & savings in your manufacturing business starts with a chat
Call Scott or Ben on 03 9955 7000 to discuss your AI Project.

Shoot us a quick question and we'll get you a quick answer

Fire away