Machine Learning: The future is now
Senior Product Manager
Virgin Media O2 Business
11th October 2022
Share this article:
Esther Petit, Senior Product Manager at Virgin Media O2 Business, discusses Artificial Intelligence and Machine Learning, and considers use cases that are being implemented by businesses right now, in 2022.
The extraordinary events of the last few years have prompted many of us to think about and discuss the future of work. Not simply where we will work, but also when and how, and what technologies and capabilities will be there to support us.
The debate has acknowledged the roles that Machine Learning might play, although often in a way that looks many years into the future, suggesting that it is some years away from being part of the everyday running of the average business.
I’d like to persuade you otherwise. In fact, powerful Machine Learning models are being deployed in almost every business sector right now. UK government research indicates that around 15% of all businesses have adopted at least one AI technology. Looking ahead, a survey of 1400 European executives found that 27% of the work of sifting large datasets for errors or actionable items will be done by machines in 2023. And research by McKinsey indicates that by 2023, intelligent chat box assistants powered by sophisticated AI engines will resolve more enquiries, faster.
At Virgin Media O2 Business, we are working with both public and private sector organisations to implement Machine Learning technologies today, which are delivering efficiencies as well as capabilities that have not been within reach until now.
So what are some Machine Learning capabilities that are becoming commonplace? Here are five significant capabilities that your organisation could benefit from right now:
Companies spend millions each year on consultancy to optimise offers and propositions, using Machine Learning technologies such as cluster analysis, to identify significant and reachable customer segments. Data analysis using Machine Learning can identify customer segments automatically, identifying trends far more effectively than any human could.
2. Predicting outcomes
Every organisation has a forecasting need, whether it is to predict temperature, footfall, revenue or energy usage. Machine Learning technologies are proving capable of highly accurate predictions, whether short or long term, as well as the ability to adjust forecasts quickly with changing data.
3. Inputs for outputs
Calculating optimum inputs for outputs is a common struggle for organisations. Machine Learning can identify patterns that might relate to the stock required to optimise revenue, staff levels to accommodate footfall or the number of trains or rolling stock needed to match passenger numbers.
4. Interpreting behaviour
Machine Learning has proven highly effective at understanding and identifying the typical features of certain behaviours. For example, how combinations of engineering work and weather patterns are likely to affect passenger journeys. Or the patterns behind the likelihood of renewing vs cancelling a contract, or of cancelling a medical appointment
5. Spotting anomalies
Machine Learning can identify a collection of patterns and behaviours that are considered normal, and so can spot outliers, or instances where the normal pattern is not met. It can be used to identify fraud, as in the previous example, but in many other situations as well, such as identifying rotten produce, or defective packaging.
We’re seeing these important applications for Machine Learning in virtually every sector right now. For example:
- In retail, deploying real-time stock availability and automated warehouse inventory management to improve the customer experience.
- In the public sector, delivering Smart City initiatives by predicting demand for public transport to optimise service provision.
- In transport and logistics, automating route optimisation for delivery drivers based on traffic data combined with drop-off scheduling.
- In manufacturing, delivering predictive maintenance using data on previous maintenance cycles and IoT inputs on current device performance
- In property, delivering informed decision-making on where to build retail or leisure facilities by combining travel and audience insights with performance data from branches.
- In financial services, improving the personalisation of offers based on insights into individual requirements and transaction data
As for us, as a service provider with over 30 million connections, we understand the value hidden in huge amounts of data. We’ve successfully navigated the challenges that Machine Learning implementations present, such as the dedicated expertise required, and managing large data sets effectively and securely. As a result, we’ve been able to benefit from the unique opportunities Machine Learning offers.
Machine Learning now plays an important role right across Virgin Media O2 Business, with multiple use cases internally and in our customer-facing activities. For example, we use Machine Learning to understand the profiles of our insurance customers, predict payment issues in our consumer contracts and target our customer messages more accurately. We also deploy Machine Learning models to protect vulnerable customers from overpaying by automatically adjusting spending caps on their accounts.
I enjoy learning about Machine Learning implementations, and talking about the new capabilities they offer. You can connect with me via LinkedIn or give us a call on 0800 955 5590 if you would like to organise a no obligation exploratory chat about how Machine Learning might be able to help your organisation.