More modern digital technologies

1 Technologies of digital transformation

The syllabus asks you to analyse digital transformation: what the underpinning technologies are, how they are evolving, what impact they have on the organisation and its operating environment – and what type of organisation they create. This chapter works through the technology toolkit named by the syllabus:

• cloud computing

• big data analytics

• process automation

• artificial intelligence – including generative AI

• data visualisation

• blockchain

• the internet of things

• mobile

• 3-D printing.

At this strategic level the interest is not in how each technology works internally, but in what it changes: cost structures, entry barriers, business models, sources of advantage and sources of risk. The chapter ends by asking what kind of organisation emerges when these technologies are combined – the digital enterprise – and Chapter 17 then builds the strategy itself.

2 Big data

2.1 What big data is

Most definitions of 'big data' describe extremely large collections of data (data sets) that can be analysed to reveal patterns, trends and associations, especially relating to human behaviour – and that are so large that conventional methods of storing and processing data cannot cope. The organised store into which such data is gathered for analysis is often called a data warehouse.

Big data is conventionally characterised by three Vs, often with a fourth added:

• Volume – a very large amount of data, more than a single computer, spreadsheet or conventional database can easily handle. The stores held by the largest retailers and technology companies are measured in petabytes (a petabyte is 10¹⁵ bytes – roughly a million typical personal computers' worth of data).

• Variety – disparate, non-uniform data of different types, sizes and sources, arriving irregularly: transactions, browsing histories, locations, text, images, voice, sensor readings. Some is internal, some external; some structured, much of it unstructured.

• Velocity – data arrives continually and often has to be processed very quickly to be useful. A price comparison texted to a customer after they have left the store is worthless.

• Veracity (the common fourth V) – is the data true and current? Analysis of wrong data produces wrong conclusions, however sophisticated the analytics.

Where does it all come from? Retailers capture every purchase (and, through websites, every product merely looked at); social media platforms hold contacts, postings, locations and photographs; mobile operators record calls, messages and a continuous location trail; internet and browser providers log every page visited and every search; banks record every payment. Increasingly, sensors (see the internet of things below) add a continuous stream of machine-generated data.

Data may be:

• Structured – held in defined fields (dates, amounts, codes) within defined records, like the entries on a bank statement. Structured data is easily interrogated with standard database query languages.

• Unstructured – with no pre-defined data model: documents, spreadsheets, presentations, emails, audio, video, social media posts, map data. A commonly quoted statistic is that around 80% of business data is unstructured, and its variety and irregularity make it much harder to store, search and analyse.

2.2 Big data analytics

The processing of big data is generally known as big data analytics. The main types are:

• Descriptive analytics – what has happened: for example, analysing sales by product, outlet, customer, time and date.

• Diagnostic analytics – why it happened: probing relationships, such as whether purchases of one product are correlated with purchases of another, or with the weather.

• Predictive analytics – what will happen: using the patterns found (regression, trend analysis, machine learning) to forecast demand or behaviour.

• Prescriptive analytics – what should be done: the system goes a stage further and recommends (or takes) the decisions that optimise the outcome – for example, redirecting stock to the outlets where predicted demand is highest.

Supporting techniques include data mining (searching data for patterns, sequences and correlations), text and voice analytics (scanning documents, emails and calls for useful signals) and statistical analytics such as the website-traffic analysis offered by tools like Google Analytics.

Strategically, the pay-offs sought are better marketing and hyperpersonalisation (Chapter 9), better customer service and relationship management, increased loyalty, sharper operational efficiency, stronger competitive position and the discovery of new sources of revenue.

2.3 Dangers of big data

• Cost – the hardware, software and skills needed are expensive (though falling in price); the investment must be justified by the value of the insights.

• Regulation and privacy – most jurisdictions now regulate the collection, storage and use of personal data. In the UK the framework is the UK GDPR and the Data Protection Act 2018, as amended by the Data (Use and Access) Act 2025; breaches can attract very large fines as well as serious reputational damage.

• Loss and theft of data – beyond regulatory penalties, organisations face civil action and lasting reputational harm if data is hacked or leaked and individuals suffer as a result.

• Veracity – incorrect or out-of-date data produces incorrect conclusions; and even correct data throws up spurious correlations that can mislead.

• Ethics and trust – detailed monitoring of employees and intrusive profiling of customers may be lawful yet still corrode trust if it goes further than people find acceptable (see Chapter 18 on ethics).

3 Blockchain technology

Blockchain technology allows data to be added to a set of records, but once added the data cannot be changed without detection.

When a piece of data (a 'block') is added, a hash – a short digital fingerprint – is calculated from it by a mathematical process. When the next block is added (like adding a link to a chain, hence 'blockchain'), the hash of the previous block is included in the new block, and a hash of the whole new block is produced. Any later alteration of an earlier block changes its hash, which then fails to agree with what was recorded in the following block, and so on down the chain: tampering is immediately visible. New records can be appended; history cannot quietly be rewritten.

The second feature is that the chain is held simultaneously on many computers in a network, which continually compare and validate their copies – a distributed ledger. A fraudster who alters the records on one computer is simply out-voted by all the others; and if one computer is destroyed, many copies survive.

Applications include property registers (an unfalsifiable history of ownership, taxes and planning permissions for a piece of land), supply-chain provenance (proving where goods have come from and how they were handled), share and contract records, and cryptocurrencies such as Bitcoin and Ethereum, which use a blockchain to record coin ownership and every transaction. Transfers made over a blockchain need no intermediary such as a bank – one reason the technology is strategically interesting: it can disintermediate the trusted third parties on which many industries' business models rest, making transactions faster and cheaper.

4 Cloud computing

With cloud computing, software and data live on remote servers run (usually) by a third-party provider, and users interact with them over the internet. Each user appears to have the program locally, but the processing happens in the provider's 'cloud' of servers. Client machines can be 'thin' – cheap and unpowerful – because they are little more than an interface, and there is only one copy of the software to update, so everyone is always on the same, current version.

The strategic attractions:

• Scalability and flexibility – computing power can be turned up and down with demand. A design firm that needs intensive processing only occasionally rents it for those moments instead of owning it permanently.

• Cost profile – heavy up-front capital expenditure on hardware is replaced by a pay-as-you-go operating cost.

• Lower barriers to entry – a start-up can rent, from day one, computing infrastructure of a scale and sophistication that once only large corporations could afford. Cloud is a great leveller, and it is the platform on which most of the other technologies in this chapter (big data analytics, AI, mobile services) are delivered.

• Location independence – staff can work from anywhere with a connection.

The corresponding risks: dependence on the provider and on communications (loss of connection means loss of processing), performance if communications are slow, and confidential data being held on third-party machines and transmitted over public systems – a security, regulatory and due-diligence issue.

5 3-D printing

3-D printing, also known as additive manufacturing, builds an object up layer by layer – plastic, metal, even concrete – rather than cutting or drilling material away as traditional subtractive processes do. Shapes can be produced that conventional machining cannot achieve, directly from computer-aided design (CAD) data, with no skilled machine operator needed to turn drawings into components.

The strategic significance lies in what becomes cheap: rapid prototyping (designs tried out before committing to full manufacture), one-off and low-volume items (spare parts for products no longer in production, custom medical implants shaped precisely from a patient's scans), and manufacturing close to the point of use rather than shipping from distant factories.

6 Process automation

Business process automation applies technology to automate and streamline processes. In manufacturing, materials can be delivered automatically to production lines where digitally controlled machinery makes whatever the schedule requires; automated warehouses pick, pack and despatch with minimal human involvement. In offices, robotic process automation (RPA) software performs rule-based clerical work – transferring data between systems, matching invoices, producing standard reports – exactly as a person would at the keyboard, but faster and without tiring.

Benefits of process automation:

• processes are performed faster and more efficiently

• more consistent results and fewer errors

• easier switching between products – reprogramming a machine rather than retraining or redeploying people

• staff are released from menial tasks to focus on more demanding work

• better communication – automated processes generate up-to-date progress and volume information automatically

• easier problem diagnosis – consistent processes make it clear when something is wrong with the process itself, rather than the noise of random human error.

7 Artificial intelligence

7.1 Machine learning

Artificial intelligence (AI) is a general term for technologies that make machines 'smart'. The engine of modern AI is machine learning (ML): software that learns patterns from data rather than being explicitly programmed with rules. Predictive (or discriminative) machine learning scores, classifies and forecasts – fraud scoring, demand forecasting, credit decisions, recommendation engines – its output being a prediction: a number, a score, a category.

An early landmark illustrates how far machine learning had come even before the current wave. In 2016 DeepMind's AlphaGo beat Lee Sedol after first learning from thousands of expert human games and then improving through self-play. AlphaGo Zero, introduced in 2017, removed the human-game training step: given the rules, it learned from self-play alone and surpassed the earlier champion-beating system. Treat both as history now: the centre of gravity has since shifted to generative AI.

Predictive ML is already deep inside business. AI grades timber arriving at sawmills and calculates how to cut each log for minimum waste; in professional firms it scans thousands of documents for the handful relevant to a case, going far beyond keyword search; auditors use it to flag transactions that look out of character for closer human attention; and natural language processing lets systems understand and respond to ordinary speech and text.

7.2 Generative AI

Generative AI (GenAI) is a class of machine-learning model trained for a different task: instead of predicting or classifying, generative models create new content – text, analysis, images, designs and computer code – in response to instructions ('prompts') written in ordinary language. The best-known examples are large language models (LLMs), trained on enormous volumes of text, which can draft, summarise, translate, explain and answer questions in fluent prose; the same family of models ('foundation models') powers image, audio and code generation, and the 'copilot' assistants now embedded in office, accounting and development software.

Strategically, generative AI matters because it automates work previously thought safely human:

• Productivity – first drafts of reports, marketing copy, analysis commentary, code and designs are produced in seconds and refined by people, compressing knowledge work the way robots compressed manual work.

• Hyperpersonalisation at scale – marketing content, product recommendations and customer-service conversations can be generated individually for each customer rather than segmented (see Chapter 9).

• New products and experiences – conversational interfaces, AI tutors and advisers, and generative design (an engineer states requirements; the system proposes designs).

• Lower cost of expertise – capabilities such as drafting, translation and basic analysis become cheap and universally available, changing the economics of service industries and professional firms.

And its strategic risks demand equal attention:

• Hallucination – generative models produce fluent, confident output that can be simply wrong. Unchecked output in customer communications, financial analysis or legal work is a serious quality and liability risk.

• Intellectual property and data governance – models are trained on material whose ownership may be contested, and confidential data typed into public AI tools may leave the organisation's control.

• Bias – models reproduce the biases of their training data, with legal and ethical consequences in areas like recruitment and lending.

• Security – new attack surfaces (manipulated prompts, poisoned data, convincing AI-generated fraud and impersonation).

• Workforce impact – roles built on drafting, summarising and routine analysis change fastest; the organisation must reskill and redesign work, a change-management challenge (Chapter 14) as much as a technical one.

Boards cannot treat any of this as an IT-department detail: choosing where to deploy AI, how to govern it and how fast to move is a strategic decision, taken up in Chapter 17.

8 The internet of things

The internet of things (IoT) is the concept of many different devices – thermostats, motion sensors, machines, vehicles, shelves – being connected to the internet and, potentially, 'talking' to each other.

Examples: in agriculture, aerial surveys assess the mineral content of grassland and GPS-guided tractors then tailor fertiliser to each area, while cattle wear collars that identify them to feeding systems dispensing the right ration for each animal. Logistics companies ship food and vaccines with temperature monitors that track and prove that limits were never breached. Production lines and parcels alike can be tracked in real time – the 'track your parcel' map is IoT at work. A warehouse shelf that senses its own weight can trigger a replenishment order as stock runs low.

Strategically, IoT gives organisations continuous, granular, real-time data about physical operations – feeding big data analytics and AI, enabling new service-based business models (see the Rolls-Royce example below), and adding new dependencies and cyber vulnerabilities: every connected device is a potential way in.

9 Data visualisation

Big data, IoT streams, trend analysis and financial models all produce data – often vast amounts. Turning it into information people can use means presenting it visually, because charts and diagrams are easier to understand and interpret, make relationships visible that tables of figures hide, prompt discoveries, and are more memorable.

Modern visualisation is interactive: dashboards let users hover for precise values, click to drill down into detail, and animate change over time. Spreadsheets provide basic charting, but specialist tools (and now AI assistants that draft charts and commentary on request) have made sophisticated visualisation a routine management expectation rather than a specialist product.

10 Mobile

Smartphones, tablets and laptops connected over WiFi and 4G/5G mobile networks put the organisation's systems wherever its people and customers are. A sales representative no longer carries printed price lists that may be out of date: current prices, product specifications, the customer's buying history and live inventory are all accessible on the spot, and the order is uploaded instantly. Airline crews carry manuals and checklists on tablets instead of kilograms of paper. For consumers, mobile is now the dominant channel through which digital services are consumed – for many businesses the app is the shop front.

11 How the technologies work together

12 The digital enterprise

The syllabus asks what type of organisation these technologies create. A digital enterprise is one whose business model and operating model are built around digital technology and data, rather than merely supported by them. Its characteristic features:

• Data-driven decisions – decisions at every level rest on analytics and experimentation, not just hierarchy and habit.

• Automated core processes – transactions, fulfilment and service run straight through with little manual intervention, so marginal costs are low and the business scales without hiring in proportion.

• Connected architecture – systems exposed through interfaces (APIs) that let partners, suppliers and complementors plug in, making the organisation easy to build ecosystems around (Chapter 17).

• Digital channels and experiences – customers are reached, served and retained primarily through digital channels, with the data each interaction generates feeding back into personalisation and product improvement.

• Agile ways of working – small cross-functional teams releasing improvements continuously, rather than long planning cycles (the leadership and culture implications are examined in Chapter 17).

Two contrasts capture the shift in business model:

Not every digital enterprise is a Silicon Valley start-up: incumbents like Rolls-Royce become digital enterprises by rebuilding their operating models around data and connectivity. How an established organisation gets from here to there – and who leads that transformation – is the subject of Chapter 17.

13 The power of social media

Social media greatly increases the speed with which accounts of real or alleged poor corporate behaviour spread – and amplifies them. If an organisation is thought to have acted unethically or irresponsibly, public pressure arrives within hours, much of it from people who have no connection with the company at all, and companies often react faster to a viral post than to a polite letter to the customer-relations department.

The influence runs both ways: organisations use social media to promote their products, their personality and their ethical commitments – from playful brand-building posts to publicising sustainability pledges. For the link between social media, reputation and ethical behaviour, see Chapter 18.

14 Test your knowledge

Two short exercises close the chapter in the online notes: ten flashcards on the terms and frameworks above, and ten practice questions with worked feedback on every option. Work through the cards first, then the questions.