Within the past year, machine learning mania has swept the business world. According to Arthur Samuel, the computer scientist who coined the term half a century ago, machine learning is defined as the subfield of computer science that employs large data sets and training algorithms to “give computers the ability to learn without being explicitly programmed.”
Many executives have an intuitive sense that machine learning will prove as important a paradigm shift as the Internet and the personal computer. A recent survey conducted by PwC indicated that 30% of business leaders believed AI would be the biggest disruption to their industry within five years. In 2016 alone, more than $5 billion of venture funding flooded into machine learning startups. The McKinsey Global Institute notes that machine learning has “broad applicability to many common work activities,” including pattern recognition, generating and understanding natural language, and process optimization.
Why Machine Learning Matters Now
The recent hype is driven by three key developments, which have reduced the barrier to entry for organizations across sector and stage that want to apply machine learning:
- More data and cheaper storage: The rise of cloud-based tools and the plummeting cost of storing data through services like Amazon Redshift mean that more data than ever is routinely generated and stored by business-critical applications.
- Open-source libraries: Widely available machine learning libraries like Google’s TensorFlow and scikit-learn make cutting edge algorithms more accessible to a wider audience of data scientists and generalist software engineers.
- Greater horsepower: The development of cloud-based platforms and custom hardware optimized for machine learning means that these applications can run faster and at lower cost, increasing their suitability for variety of business needs.
In the abstract, there is compelling evidence to invest in machine learning. But how do organizations really use this technology? In what ways is machine learning deployed today to help companies create value, cut costs and drive ROI?
In this article, we share case studies illustrating how companies of all sizes employ machine learning to address five key business cases: user acquisition, customer support, forecasting, fraud prevention, and people management.
1. User Acquisition
In broad strokes, the customer acquisition funnel for a typical consumer or enterprise business has three stages: segmenting your customer base to understand and address their needs, engaging them with the right messaging at the right time, and converting them into users of your product.
Machine learning has seen wide use by startups and major corporations alike across the entire user acquisition funnel. Amazon is a key example here—in his 2017 letter to shareholders, CEO Jeff Bezos remarked on the ways that machine learning contributes to the Amazon.com experience “beneath the surface” by powering product and deal recommendations based on user preferences. But segmenting users and showing them relevant products is only the first step: many retailers use machine learning to adjust branding, copy and promotional pricing on the fly to maximize the likelihood of a sale for any given customer.
On the enterprise front, Salesforce recently launched Einstein, a product that examines CRM data to provide tailored recommendations to increase the chance that a particular prospect will convert from a sales pitch, going so far as to suggest the right time to send an email.
2. Customer Support
Of course, acquiring customers is only the first step. Whether for ecommerce or the enterprise, retaining users and limiting churn requires providing timely and effective customer support.
Dozens of brands now make use of machine learning to improve the customer support experience. For example, Brazilian supermarket Ocado used Google machine learning APIs to build a custom system that measures the sentiment of customer support inquiries and moves negative responses to the top of the support cue. The result is that Ocado responds to urgent messages four times faster, creating a valuable opportunity to win back customers at high risk of becoming detractors.
More recently, conversational “bots” are now triaging support requests without help from a human operator—using machine-powered natural language to deliver a first response that can fulfill routine requests like issuing return labels. Besides reducing support costs by up to 30%, chatbots can boost customer satisfaction by responding faster, and the scope of their capabilities will grow as their comprehension skills improve. With a staggering 44% of U.S. consumers preferring to interact with chatbots over humans, consumer-facing enterprises that invest in machine learning will have a tremendous advantage.
In the back office, a wide variety of organizations are starting to use machine learning to build more robust, granular and accurate forecasting models.
In 2016, Walmart ran a competition on the data science recruiting platform Kaggle, asking applicants to use historical data from 45 stores to build a model that forecasted sales by department for each store. Insurance giant AIG has assembled a 125 person data science team to build machine learning models, with the goal of improving the company’s ability to anticipate claims and predict outcomes.
Even the global eyewear conglomerate Luxottica puts machine learning to work forecasting demand: it adds 2000 new styles to its collection every year, and uses machine learning and data from past launches to predict sales performance.
4. Security and Fraud Detection
In 2016, fraud cost the average ecommerce retailer over 7% of total revenue. Salaries for fraud management employees, chargebacks, and legitimate transactions that are denied due to false positives all contribute to this expense.
Machine learning is starting to bear out its potential as a powerful tool to intelligently monitor millions of transactions in real-time, reducing waste from fraud. PayPal is a leader in this arena: they have used open-source tools and their vast trove of transaction data to build an artificial intelligence engine from scratch, with the key goal of reducing the number of false alarms produced by their older fraud models.
Humans are still in the loop to train the model and sort out ambiguities, but the initial result has been astounding: since implementing their new model, PayPal has cut its false positive rate by half. For companies seeking a white-glove solution, startups like Sift Science can consume a business’s data and apply fraud signals from their entire network of enterprise customers, ensuring that the latest techniques of fraudsters are swiftly caught.
5. People Management
Hiring, managing and retaining high-quality people is the root of all business competencies. One of the most onerous parts of hiring is filtering hundreds or thousands of resumes to assemble a shortlist for interviews; over half of recruiters say this is the most difficult part of their job. This problem is being addressed by startups like Restless Bandit, which makes a candidate management system used by companies like Adidas and Macy’s to filter resumes based on decisions that hiring managers have made in the past.
Crucially, these algorithms can be trained to ignore unconscious human biases and even flag biased language in job descriptions—meaning that machine learning has the potential to identify high-performing, diverse candidates who may be overlooked by human recruiters on a first pass. On the retention front, machine learning can augment the mentorship of great managers and help employees perform better by generating specific and unbiased career advice, based on past employees with similar profiles.
With a staggering 44% of U.S. consumers preferring to interact with chatbots over humans, consumer-facing enterprises that invest in machine learning will have a tremendous advantage.
The Impact of Machine Learning Will Grow
In this article, we have reviewed some of the most significant ways that machine learning can create direct and immediate value for a variety of organizations. It would be a mistake to view machine learning as some kind of corporate panacea—ultimately, the performance of a machine learning system is only as good as the data on which it is trained, and an enterprise’s key decisions are often “edge cases” that require a measure of human judgement and anecdotal experience to assess.
Instead of being dazzled by the abstract potential of machine learning, executives should approach the question of investing in this technology by taking stock of their core business challenges and matching them against the key capability of machine learning: drawing sense and meaning from a ton of data. Given the diversity of case studies above, the odds that machine learning techniques can help may be greater than you expect.