AI tackles profound technical challenges

AI is significant because it successfully tackles a profound set of technical challenges. Increasingly, human capabilities – understanding, reasoning, planning, communication and perception – can be undertaken by software, at scale and at low cost. General analytical tasks, including finding patterns in data, that have been performed by software for many years can also be performed more effectively using AI.

Together, these capabilities create new opportunities in most business processes and consumer applications.

AI research is focused on five fields

Since its inception in the 1950s, AI research has focused on five fields of enquiry:

1. Knowledge: The ability to represent knowledge about the world. For software to possess knowledge, it must understand that: certain entities, facts and situations exist in the world; these entities have properties (including relationships to one another); and these entities and properties can be categorised.

2. Reasoning: The ability to solve problems through logical reasoning. To reason is to apply logic to derive beliefs, related ideas and conclusions from information. Reasoning may be deductive (derive specific conclusions from general premises believed to be true), inductive (infer general conclusions from specific premises) or abductive (seek the simplest and most likely explanation for an observation).

3. Planning: The ability to set and achieve goals. For software to be able to plan, it must be capable of specifying a future, desirable state of the world and a sequence of actions enabling progress towards it.

4. Communication: The ability to understand written and spoken language. To communicate with people, software must have the ability to identify, understand and synthesise written or spoken human language.

5. Perception: The ability to make deductions about the world based on sensory input. To perceive, software must be able to organise, identify and interpret visual images, sounds and other sensory inputs.

Progress in AI has unlocked new possibilities

Because most business processes and consumer applications involve knowledge management, reasoning, planning, communication or perception, progress in AI has unlocked significant new capabilities.

In the following chapter, we describe specific AI use cases in eight sectors.

The applications of AI in industry are numerous and tangible

AI is not a solution seeking a problem; it is a tangible set of capabilities unlocking revenue growth and cost savings. The capabilities of AI – its power to incorporate broader data sets into analyses, identify concepts and patterns in data better than rules-based systems, and enable human-to-machine conversation – have applications in all sectors and numerous business processes. In approximately 60% of occupations, at least 30% of constituent activities are technically automatable by adapting currently proven AI technologies (McKinsey Global Institute). As such, AI is a key ‘enabling technology’.

Data-centric sectors will see the greatest impact

AI is being deployed in all sectors and to a wide variety of business processes. However, AI will have more numerous applications and greater impact in certain sectors.

AI’s impact will be greatest in sectors in which a large proportion of time is spent collecting or synthesising data, or undertaking predictable physical work. In several sectors (Fig. 17), professionals spend one third or more of their time on the above (McKinsey, Julius Baer).

These sectors include:

• Finance and insurance (50% of time)
• Retail, transport and trade (40% of time)
• Professional services (37% of time)
• Manufacturing (33% of time)
• Healthcare (33% of time)

Applications will be more limited in sectors in which data synthesis and processing activities are limited, or in which the majority of people’s time is spent managing others or undertaking unpredictable physical work. Occupations such as management and teaching will be more resilient to AI in the medium term.

Core use cases vary by sector

Use cases for AI are proliferating as understanding of the technology improves. Below, we describe 31 core AI use cases in eight sectors: asset management, healthcare, insurance, law & compliance, manufacturing, retail, transport and utilities.

Asset management

AI’s ability to extract content from unstructured data using natural language processing, find subtle patterns in disparate data sets, and enable machine-to-human communication via chatbots, has multiple applications in asset management. Core use cases include investment strategy, portfolio construction, risk management and client service.

By augmenting or automating many of an asset manager’s tasks, AI enables asset managers to deliver to the mass affluent a degree of personalisation and service quality previously reserved for high net worth clients. Additionally, AI can decrease the cost of portfolio construction while improving quality – the era of the ‘robo-advisor’.

Investment strategy: AI can improve a firm’s investment strategy by synthesising its research and data, and incorporating broader data sets including unstructured information. Superior pattern recognition can then deliver better multi-objective optimisation. AI can balance a diverse range of inter-connected objectives (including fund deployment, risk and profitability) to enhance returns more effectively than rules-based systems.

Portfolio construction: AI tools can augment, and increasingly automate, an asset manager’s process of portfolio construction. AI – ‘robo-advisors’ – can analyse a client’s goals, and within a firm’s investment rules develop personalised, optimised portfolios at low cost and high speed.

Risk management: AI can improve risk management by incorporating broader data sets and improving analytical processing. 90% of data generated today is unstructured information, stored outside traditional databases (International Data Group). Natural language processing enables additional data sets to be incorporated into firms’ analyses. Other AI techniques, including deep learning, then enable patterns in data to be identified with greater granularity and confidence. Together, these capabilities enable risks to be identified and quantified more effectively.

Client service: Chatbot interfaces are being applied within and beyond asset management firms. Deployed in clientfacing channels, natural language systems enable client enrolment, support and self-service. Embedded in internal tools, chatbots let account managers query client details and  understand developments relevant to a client’s portfolio in seconds instead of minutes. Fewer account managers can then provide a higher quality service to a greater number of clients.

Healthcare

In the next decade, AI can unlock a paradigm shift in healthcare to improve patient care and process efficiency. Automated diagnosis was an early use case for rudimentary AI in the 1980s. ‘Expert systems’ mimicked human approaches to diagnosis, applying rules-based inferences to bodies of knowledge. Modern AI, particularly deep learning, is more effective and applicable to a wider range of processes. Key use cases include diagnosis, drug discovery and patient monitoring.

Diagnosis: Deep learning systems can replace complex, human-coded sets of probabilistic rules and identify subtle correlations between vast, multi-variate data sets to deliver scalable, automated diagnosis. While systems are nascent, accuracy is improving rapidly. Separately, computer vision solutions powered by deep learning are transforming diagnostic imaging. While human radiologists require extensive expertise and years of training to identify abnormalities in magnetic resonance images and ultrasounds, deep learning systems trained on large data sets deliver impressive results. Diagnostic imaging, powered by deep learning, now offers human-level accuracy and high speed in select contexts.

Drug discovery: Today’s drug discovery process is lengthy, averaging 12 years to market (California Biomedical Research Association). Expense and uncertainty are also prohibitive; drug development costs an average of $359m and just 2% of US preclinical drugs are approved for human use (California Biomedical Research Association). AI is being applied to multiple stages of the drug development process to accelerate time to market and reduce uncertainty. AI is being applied to synthesise information and offer hypotheses from the 10,000 research papers published daily, predict how compounds will behave from an earlier stage of the testing process, and identify patients for clinical trials.

Insurance

While the fundamentals of insurance – customer prospecting, risk assessment, claims processing and fraud detection – have remained unchanged, modern AI can improve every stage in the insurance process to deliver efficiency savings and improved customer experience. By identifying patterns in data better than rules-based systems, AI can improve and accelerate decision-making and claims processing, reduce fraud and automate a large proportion of customer service enquiries.

Risk assessment: AI can gather information from broader data sets, including web and social media profiles, to compile richer customer information and inform risk assessment. AI can then assess the risk of individual policies more accurately than rulesbased systems, by detecting non-linear patterns in multi-variate data sets and making more accurate projections.

Claims processing: AI can reduce time-to-quote, time-to-claim and claims processing costs for consumers and insurers. To accelerate claims processing, AI systems can automatically extract and classify structured and unstructured information from insurance policies and claim forms. By analysing images of damaged assets, computer vision systems can automatically classify claims. Through improved pattern recognition applied to prior cases, AI can also predict settlement costs. Algorithms using deep learning are effective for image analysis, while Bayesian (probability-based) AI is useful for predicting settlement costs.

Fraud detection: Insurance fraud costs UK insurers £1.3bn annually and adds £50 to the average policyholder’s annual bill (Association of British Insurers). UK insurers invest over £200m annually to tackle the challenge (Association of British Insurers). Fraud detection algorithms enhanced with AI can identify
fraudulent transactions, while reducing false positives, more effectively than traditional approaches.

Customer service: Chatbot interfaces integrated with insurers’ databases can use natural language processing to offer 24/7 product information and answers to policyholders’ enquiries in a scalable, inexpensive and personalised channel.

Law and Compliance

AI’s abilities to process language in documents, synthesise knowledge and automate reasoning have broad application in the legal services and compliance sector. With junior lawyers spending a high proportion of their time accessing and collating information, scope for augmentation and automation is considerable. Key AI use cases include identifying relevant case law, processing documents for discovery and due diligence, and informing litigation strategy. In October 2018, Harvard Law School Library advanced its ‘Caselaw Access Project’ by releasing over 40 million pages of digitised legal information, including every reported state and federal legal case in the United States from the 1600s to the summer of 2017 – providing extensive further data to train AI systems.

Regarding compliance, costs have grown significantly since 2008 – particularly for financial services firms. With rules-based software poorly suited to catching infractions, banks have invested in additional compliance personnel. Citi, while reducing its global headcount 32% between 2008 and 2016, doubled its regulatory and compliance staff to 29,000 – over 13% of it workforce (Citi). AI’s ability to learn patterns of behaviour over time, and highlight unusual activity in real-time, offers greater scalability at lower cost.

Case law, discovery and due diligence: Natural language processing AI can identify, classify and utilise content from databases and unstructured documents at scale and speed, saving legal firms time and cost for routine document review. Use cases include sourcing and ranking relevant case law and identifying key documents in due diligence and discovery processes. With a merger and acquisition data room containing an average of 34,000 pages for review (Luminance), AI can increase business velocity and reduce costs.

Litigation strategy: AI can analyse past judgements at greater speed, granularity and subtlety than has been possible previously. By anticipating the probability of different outcomes, lawyers can inform and enhance their strategic decision-making. In high volume areas, such as personal injury, software can help a firm decide whether to accept a case. In high value areas, including corporate litigation, software can suggest the probability of a particular outcome based on juries’ prior behaviour and opposing lawyers’ tendency to settle or proceed to trial.

Compliance: Preventing accidental or deliberate breaches of policy, from the theft of sensitive data to accidentally misaddressing an email containing a customer database, is challenging for rules-based systems. By learning the habits of users over time, AI systems can flag potential compliance breaches in real-time, before they occur, with sufficient accuracy to enable broad deployment.

Manufacturing

AI can significantly improve manufacturers’ efficiency and profitability. Overall Equipment Effectiveness (OEE), a measure of manufacturers’ productivity  relative to potential, varies widely by industry, from 75%-91% (LNS Research). The performance of companies within the same industry also varies widely, offering scope for competitive advantage. AI can boost OEE and profitability by predicting equipment failure (to reduce unplanned downtime), improving  assets’ operational efficiency, and reducing utility supply costs.

Predictive maintenance: The failure of production assets is costly; one hour of unplanned downtime on an automotive assembly line can cost a manufacturer £1.5m (MMC Ventures). AI can identify subtle patterns in data from vibration, temperature, pressure and other sensors to identify leading indicators of equipment failure. By predicting more accurately which components are likely to fail, and when, parts can be proactively replaced to prevent failures and save money.

Asset performance: AI can improve the operation of high value assets, including gas and wind turbines, to optimise yield. Rules-based programs deliver limited results when applied to complex tasks, such as tuning fuel valves on a gas turbine to optimise combustion while reducing wear and emissions. Applying neural networks to optimise the turbine inputs can improve results by 20% or more.

Utility optimisation: Optimising the purchase and consumption of utilities, such as power and water, according to real-time demands on a factory floor is too challenging and variable to manage using rules-based software. AI enables companies to anticipate, and align, utility consumption with process requirements in realtime, lowering utility consumption by 5% or more.

Retail

E-commerce, now 17% of UK retail sales and growing (eMarketer), has transformed the quantity, breadth and granularity of data available to retailers. Retailers that turn data into insight can increase competitive advantage by engaging, monetising and retaining customers more effectively. Every stage of a retailer’s customer journey – from lead generation and content selection to price optimisation and churn prediction – can be improved by AI algorithms that ingest richer data sets and identify patterns in them better than rules-based systems. By enabling analytics at the ‘per-customer’ level, AI is introducing the era of retail personalisation. Leaders enjoy competitive advantage; 75% of Netflix users select films recommended to them by the Company’s AI algorithm (Netflix).

Customer segmentation: Limitations in available data, and the linear analysis of information, inhibit the ability of traditional customer segmentation software to identify desirable customer attributes. Deep learning algorithms enable natural language processing, which enables retailers to access additional data sets including social media data. Deep learning algorithms also offer more granular analysis than rules-based systems, to optimise segmentation, channel selection and messaging.

Content personalisation: Most content presented to online shoppers is irrelevant or poorly suited to users’ preferences, reducing conversion to an average of 1.0% on smartphones and 2.8% on desktops (Adobe). As with customer segmentation, AI offers additional unstructured data sets for analysis, and improved multivariate analysis to identify more subtle correlations than rules-based systems can detect. When Netflix recommends content to a user, in addition to analysing a user’s actions, ratings and searches, the Company’s AI algorithm considers social media data and meta-data from third parties. The Company is now analysing images from content, including colour palette and scenery, for deeper personalisation.

Price optimisation: A 1% change in price provides, on average, a 10% change in profitability (BlueYonder). The smaller a company’s margins, the greater the impact. Willingness to pay is a key determinant for price. AI enables price optimisation that is more sophisticated than traditional ‘cost plus’, ‘relative-to-competitors’ or ‘odd pricing’ (£0.99) models. By identifying correlations within and between data sets, AI can better optimise for factors including price elasticity, revenue, profit, product availability and phases in a product’s lifecycle (introduction or end-of-life).

Churn prediction: Traditional programs struggle to incorporate new sources of information, maximise the value from multi-variate data sets or offer granular recommendations. AI-powered churn prediction can identify leading indicators of churn more effectively, and improve remediation by predicting more accurately the format and content of successful interventions.

Transport

The transport sector will be transformed by AI. Breakthroughs in computer vision are enabling the age of autonomous vehicles – self-driving cars, buses and trucks. The implications, from shifts in sector value chains to new business models, will be profound (see Chapter 8). In addition to enabling autonomy, AI can be applied to the many prediction and optimisation challenges – from congestion modelling to fleet management – at the core of today’s logistics networks.

Autonomous vehicles: AI computer vision systems enable vehicles to sense and identify the physical features and dynamics of their environment, from road lanes to pedestrians and traffic lights, with a high degree of accuracy. Combined with AI data processing and planning algorithms, AI is enabling the age of autonomous transport. Cars, buses and trucks will be able to operate and guide themselves, without human involvement. SAE International, a US-based global professional association and standards body, has identified five degrees of vehicle autonomy, from Level 0 (no automation) to Level 5 (full automation; no requirement for human control).

Select companies, including Google, intend to release vehicles offering Level 5 automation. Challenged by the autonomous vehicle programmes of Google, Uber and Tesla, incumbent manufacturers are accelerating their own initiatives by increasing investment and making acquisitions. Ford intends to deliver high-volume availability of at least a Level 4 autonomous vehicle by 2021. In October 2018, in the UK, private hire firm Addison Lee announced its intention to deploy self-driving cars in London by 2021, by partnering with UK autonomy company Oxbotica.

Infrastructure and system optimisation: AI’s abilities to detect patterns and optimise complex data are being applied to traffic, congestion and infrastructure challenges in transport systems. Predicting traffic flows, or modelling the deterioration of transport infrastructure, are difficult because inputs are complex (combining traffic, construction and environmental data) and because the relationships between inputs and outputs are non-linear (Transportation Research Circular). In these contexts, machine learning and deep learning systems are well suited to deliver better results than rules-based systems.

Fleet management: Transportation fleets are pervasive, from the logistics networks that underpin the economy to taxi fleets and food delivery services that provide point-topoint convenience. AI can optimise pick-ups, route planning and delivery scheduling to maximise asset utilisation, while considering economic, social and environmental impacts.

Control applications: Machine learning systems are well suited to the numerous prediction and optimisation challenges presented by air traffic control, vehicle traffic signalling, and train control.

Utilities

Information processing will become critical to utility companies, and their business models, as the utility sector undergoes a greater change in the next 25 years than it has during the previous 150. ‘Prosumers’ – consumers who also own capacity for energy production – will require integration into the energy market. By processing data more intelligently, AI will be a significant value driver in this transition. AI use cases for utility companies are varied, from demand optimisation and security to customer experience.

The foundations for AI adoption in the utilities sector are robust. 67% of utility companies – a higher proportion than in any other sector – use ‘internet of things’ (IoT) technologies such as sensors (Gartner). Further, compared with peers in other sectors, utility CIOs have a stronger focus on cost reduction,
managing geographically dispersed assets and security.

Supply management: AI algorithms can predict changes in supply, including those caused by the intermittency of renewable resources, more effectively than rules-based systems – enabling smaller reserves and greater cost savings. AI solutions can also optimise supply networks, which are becoming increasingly complex as consumers deploy sources of renewable energy that contribute energy back to the National Grid.

Demand optimisation: By identifying detailed patterns in consumer behaviour, AI algorithms can move consumption of energy from periods of peak use and high prices to times of lower demand and cost.

Security: Rules-based systems struggle to deliver system security given the continually evolving nature of security threats. By identifying abnormal patterns in network behaviour, deep learning systems can identify breaches in network security that elude traditional programs.

Customer experience: Chatbots, which offer natural language conversations powered by deep learning algorithms, offer consumers self-service account administration, product information and customer service.

In Chapter 8, we explore the profound implications of the proliferation of AI across multiple sectors.