As part of Information Age’s blockchain and emerging technology month, we explore the transformational impact of advanced AI and analytics. AI and advanced analytics can have a transformational impact on every aspect of a business, from the contact centre or supply chain to the overall business strategy. With the new challenges caused by coronavirus, companies are in a growing need of more advice, more data and visibility to minimise the business impact of the virus. However, long before the disruption caused by Covid-19, data was recognised as an essential asset in delivering improved customer service. And yet, businesses of all sizes have continued to struggle with gaining more tangible value from their vast hoards of data to improve the employee and customer experience. Data silos, creaking legacy systems and fast-paced, agile competitors have made the need to harness an organisations data to drive value of paramount importance. The challenge is huge and many traditional and new companies are waking up to the use of the partner ecosystem and the need to utilise various technologies, like AI and advanced analytics, to stave off disruption and innovate by taking advantage of data. From adopting industry standards to the use of graph databases and a real-life use case of AI and advanced analytics in action, six experts explore the transformational impact of AI and advanced analytics, while explaining how to implement the technologies.
1. Align data strategy to business goals
Patrick Smith, Field CTO EMEA at Pure Storage, understands the value of data. It’s the “most valuable form of modern currency,” he says. However, he points out that “vast swathes of business data is only actionable if it can be processed, read and understood fast. In this sense, advanced analytics is data’s unsung hero. It does all the heavy-lifting and underpins business transformation efforts and helps companies both big and small to increase their results and performance.” Despite this knowledge, Smith highlights that most organisations lack the infrastructure and analytical software or the know-how to implement AI and advanced analytics effectively. He explains that to overcome this, companies “must be laser-focused on aligning their data strategy to their business goals, and work with technology partners to provide a modern data experience based on infrastructure that is lightning fast, scale-out and easy to use.”
Data and analytics predictions and priorities for 2020
2. Moving away from manual process
For the last decade, business intelligence has been used to gain insight from historical data, but until recently, these analytical techniques have been mainly manual. This is changing and Wayne Butterfield, director at global technology research and advisory firm, ISG, explains that business leaders are welcoming “the promise of artificial intelligence (AI) to both remove the manual process and improve the quality of insight.” He says: “Data-driven insights — using historical data to predict future outcomes — combine data, advanced analytics and AI to transform decision making, based on predictive insights in areas like revenue, demand and supply. “It’s still early days, but auto machine learning (AutoML) technologies are lowering the barrier to entry for organisations that may not have large teams of data scientists, but that still see the value in looking forward and not backwards with their data.” Pointing to AutoML tools, like Kortical.io and Data Robot, Butterfield explains that these are “becoming more popular in automation centres of excellence, as advanced AI models are plunged into the relatively simple robotic process automation-type processes, to take action based on these predictions.”
3. A complete view
Kerrie Heath, European sales director, AI at OpenText, says that extracting value from data shouldn’t be a daunting task. By adopting advanced AI-powered analytics, organisation’s can drive value real-time “and deliver it in a visual, interactive format that lets users easily make predictions about products, topics, events, trends, and even themes and emotions,” she says. “Only with a complete view of this unstructured data and combining it with structured data from enterprise systems in real-time, will organisations be able to analyse, understand and manage their enterprise digital ecosystem more efficiently. In turn, organisations are providing themselves with the tools to ensure and enforce data governance,” adds Heath.
The hardest part of AI and analytics is not AI, it’s data management
4. Industry standards needed
Alejandro Saucedo, engineering director at Seldon, believes the implementation of advanced AI and analytics is having a tremendous impact on society. He says: “Both of these technologies lead to massive surges in productivity, and huge reductions in costs – both opportunity costs, and actual costs. Efficiency improvements from advanced AI will transform certain sectors over the next few years, notably transport, energy, and infrastructure.” However, Saucedo points out that if not implemented properly, AI can bring about undesirable outcomes for organisations, particularly when it comes to compromised cybersecurity, privacy, and trust. “To optimally implement AI and ensure it provides a net gain for our economy and society, we need to develop general and industry-specific standards, and fit-for-purpose regulatory frameworks. Transparent and enforceable frameworks are key, and we need to guarantee that relevant experts, technical and non-technical, are continuously involved in developing and updating them,” he advises.
“AI isn’t able to tell the future — even the most advanced AI couldn’t have predicted Covid-19 or its effects on the world, for example — but today’s AI models will be able to use the data from this time, including the impact that COVID-19 has had, to inform future predictions” — Butterfield.
5. Graph databases
Amy Hodler, analytics and AI programme manager at graph database firm Neo4j, says that “the logical extension of analytics is to use the relationships and network structures that are held in all data and are proven to be extremely predictive. This will transform analytics and AI as connectivity-based learning is necessary to address complex questions, including those about system dynamics and group behaviour, with less data. “Businesses can leverage connected data insights held in a graph database with efficiency and flexibility otherwise unattainable using a relational database. Because a graph database is built to preserve and compute over relationships, it enables valuable and often nuanced predictions, such as pinpointing interactions that indicate fraud, identifying similar entities or individuals, finding the most influential elements in a patient or customer journey, or even ameliorate the spread of IT or phone outages.” She continues: “Data scientists gain a force multiplier when they use graph algorithms to understand the natural shape of complex systems through data patterns and increase predictive accuracy. When used in a framework that automatically transforms a stored graph into a computational graph, they benefit from a flexible data structure that provides better predictions, more automation and contextually responsive AI.”
6. AI and advanced analytics in practice: the insurance industry
Paul Donnelly, executive vice president of EMEA at Munich Re Automation Solutions, explains how AI and advanced analytics are used in his industry, life insurance. He says: “Insurance is rife with manual processes and back office procedural steps, leading to poor customer experiences. And while purchasing life insurance isn’t something we look forward to anyway, complex processes certainly don’t help entice modern digital-savvy customers. “This is where AI and data analytics come in. Such advanced technology optimises the end-customer’s journey for many reasons. For example, harnessing AI techniques means that we can bypass the need to ask customers endless, repeated, personal questions and instead route them through the questions which are relevant to them. Because in a world where we can easily buy most products we want in minutes with a few clicks, a drawn-out life insurance process simply is not appealing. “Furthermore, analytics allow insurers to take advantage of vast amounts of applicant data and transform it into actionable insight. These insights allow insurers to amend underwriting rules in real-time, resulting in technologies that design, evolve and streamline interview processes for customer convenience and faster time to underwrite the customer. An insurer that doesn’t do so is being careless with its customers’ time.