Reinvention is not a novel concept; it has been the lifeblood of enduring organizations for decades. History presents us with numerous examples in which transforming ways of working or finding new sources of value served as a deciding factor between an organization thriving and faltering.
Netflix’s pivot from mailing DVDs etched with existing content to delivering streaming services and original productions often comes to mind as a quintessential case in point. However, there are many long-lasting organizations that have pulled the reinvention lever, with technology also providing the catalyst—think Samsung moving from grocery to textiles to electronics, or Amazon evolving from bookseller to e-commerce company to cloud services provider. Their journeys and many others offer inspiration to organizations that find themselves facing an urgent and undeniable reinvention imperative.
This year’s CEO Survey shows many organizations feeling such pressure, with nearly half of respondents saying they believe their business will no longer be viable in ten years if it stays on its current path. Executives see key drivers, from regulation to evolving customer preferences, threatening their current business’s viability more than ever before.
But technology stands out among them as the strongest reinvention force, with 56% of CEOs saying it will significantly change the way their organization creates, delivers, and captures value over the next three years.
It’s safe to say that this is at least in part due to the arrival of generative AI, a seminal innovation on par with the printing press, electricity, and the internet. This technology is poised to completely transform the nature of work and provide a fast track for new, AI-native organizations to disrupt existing markets.
CEOs recognize both the threat and opportunity, with 70% saying they see generative AI impacting their business model in the next three years—a number that increases to 89% among those at organizations who’ve already put the technology to work. This suggests that seeing GenAI in action clearly demonstrates its transformative potential.
Considering this, it’s baffling to see so many business leaders focused heavily on two topics: the technology’s cost impacts and narrow use cases. Focusing too much on productivity and incremental use-case gains understates both the bigger threat of disruption and the opportunity to reimagine workflows, business processes, and even business models.
That’s not to say organizations should ignore the productivity play. Efficiency gains will eventually become table stakes, so organizations can ill-afford to eschew them. But disruption at the business or operating model level will occur, albeit unevenly across the value chain in different sectors. This will shape each organization’s cadence in applying generative AI for efficiencies and reinvention alongside other deciding factors, such as organizational readiness, regulation, and the technical complexity of executing transformational opportunities.
In this article, we’ll share an initial analysis of how these dynamics could unfold in various industries to help organizations plot their generative AI path, to be followed by future articles that dive deep into specific sectors. We then review our value-driven flywheel implementation strategy to show how its focus on pattern recognition can enable organizations in any industry to both capture productivity benefits and drive toward GenAI-enabled reinvention—before unencumbered AI natives swoop in and upend a sector.
Which industries will GenAI impact most?
Our analysis suggests varying levels of impact potential for industries from GenAI adoption. There will, of course, also be plenty of variability among organizations, which is reflected in the range of potential operating margin uplift per sector. This range is calculated by assessing the cumulative use-case impact on benchmarked profits and losses for each sector (which are based on current operating models), and factors in GenAI capabilities today and those on the near-term horizon, and does not include the build or run costs to develop or implement generative AI, given the wide variability in these values from company to company.
Unsurprisingly, we expect organizations in the tech industry to see the highest gains in the form of a potential 19 percentage point uplift in operating margin on average. Within these organizations, GenAI can expand margins by, for example, speeding software development—responsible for as much as 20% of their costs—in numerous ways, including by assisting developers with coding, creating technical documentation, and extracting data features.
Software providers are also already bolstering their existing product offerings with GenAI. As just a few of many examples, Microsoft, SAP, and others have added GenAI-powered copilots into their enterprise software to help with everything from summarizing documents to improving employee experience; Adobe enhanced its creative suite with image generation and editing capabilities; and numerous providers, such as Amazon, have created GenAI tools that help IT professionals learn about and select the right software for their needs.
Perhaps more exciting is the innovation that’s on the horizon for both software providers and companies in every sector. Because GenAI-powered coding assistance speeds up the development process, it provides more time for software innovation. As a result, an explosion of new software—some enhancing and automating the functions of today, and others offering entirely new applications—is just around the corner.
Luxury goods could also see impressive upside from GenAI adoption (an average of 14.5 percentage points). The technology offers the ability to create hyper-personalized advertising, customer experiences, and even product designs. We’re already seeing organizations dive in, with luxury brands like Gucci using generative AI to create new product designs by combing current trends; Bally using it to deploy virtual shoppers that help them predict customer behaviors; and others such as Moncler, Zegna, and Valentino tapping it for help with advertising campaign imagery.
Although our analysis projects that these and other industries stand to gain the most on a percentage-point basis, the possible margin uplift for even those sectors in the lower benefit ranges remains substantial, considering the razor-thin, single-digit margins prevalent among them today.
Consider the transport and logistics sector, for instance, where standard freight carriers often see margins as low as 2%. With a potential 1.8 percentage-point average uplift, some organizations in the sector could see their margins nearly double. Use cases for the sector range from route optimization in outbound logistics to optimized real-time pricing in sales and automated financial reporting in the back office.
Of course, much of the low-hanging value gains available today will erode over time. They’re largely derived from efficiencies afforded by GenAI and from applying the technology to an organization as it operates today. Such use cases will become the norm when every player in the market starts leveraging similar generative AI technologies in similar ways and customer expectations inevitably advance. That’s why organizations will want to use today’s work integrating the technology as the foundation for the more transformative applications that will move businesses toward new operating models—and into new markets.
The state of play: A view through the industry lens
Industry dynamics feature a starring role in determining the speed with which organizations are likely to capitalize on generative AI opportunities, particularly those that spark significant reinvention, whether that be a dramatic business model shift or a more modest reimagination of a specific function or key business process. This will be based broadly on the level of GenAI-driven disruption potential within an industry and the ease by which organizations can integrate the technology. Several factors contribute to each of these variables.
Level of disruption. Although impacts from generative AI are set to be universally transformative, its potential for disruptive effects will vary across industries. The nature of disruption will also vary, taking the form of business model, operational, and/or competitive disruption.
In some sectors, the threat to business models—for example, the fundamental reshaping of product offerings, pricing models, and customer engagement—will be high. Consider the entertainment industry, for example, where the rapid rise of GenAI-powered content creation tools is driving a complete rethink of content development, distribution, and pricing.
Other industries, such as consumer packaged goods, might find the most significant disruption taking place behind the scenes, in the streamlining of operations and the upskilling of their workforce.
On the competitive front, the early adoption of generative AI can provide a significant leg up in some industries, like professional and legal services, where capturing market share from loyal clients (who often engage multiple firms in this space) is critical. In others, such as telecommunications, barriers for new market entrants will remain high, creating a relatively lower first-mover advantage for some GenAI endeavors, though not all.
Ease of adoption. Adopting generative AI is not a straightforward process, and the path to integration will, in cases, be industry-specific. For certain sectors, customizing AI to suit a business’s unique requirements can represent a major undertaking, heavily dependent on the availability, volume, and intricacy of relevant data, as well as a significant degree of model customization.
Consider insurance, for example, which harbors vast amounts of unstructured data that require considerable, though certainly not prohibitive, computational resources to crunch. A high degree of effort is required to apply industry- and company-specific policies, documents, and data to enable personalized policy generation and claims processing.
The ease of adoption is also influenced by other internal and external factors. The readiness of an industry’s workforce and culture to embrace new technology is one critical internal factor (though naturally with great variability among organizations within an industry). Some industries will see more disruption in the nature of worker tasks and roles than others; for example, sectors like retail and financial services, in which customer service, a function already seeing significant transformation from generative AI, features prominently. Additionally, the ability to engage in responsible AI practices—proper governance, transparency, fairness, security, and the like—presents a varying degree of difficulty to industries based on factors such as customer expectations, the sensitivity of the requisite data, and the current level of progress specific organizations have made toward embedding responsible AI practices into the fabric of their organization. External factors, such as regulatory constraints or enablers, play a role as well.
A look at 22 industries and four adoption scenarios
We plotted 22 industries against the two variables of level of disruption and ease of adoption. As with projected sector impacts from GenAI, there will be plenty of variability among organizations. And, of course, the emergence of AI-native disruptors cannot be reliably predicted. But, broadly speaking, we foresee the march toward generative AI adoption taking shape in four different ways—with companies being disruptors, trailblazers, multitaskers, or streamliners, depending on which quadrant their industry falls into.
Trailblazers (high ease of adoption and high level of disruption). Trailblazers are in a prime position to leverage generative AI for transformative applications in the nearer term. These sectors are already seeing the emergence of compelling generative AI capabilities that will enable agile players to rapidly deploy innovative business models that will disrupt the status quo.
The entertainment industry, for example, is on the precipice of seismic shifts in the way music, movies, and series are created, thanks to ongoing and rapid improvements in generative music and video tools offered by both incumbents and newcomers. Market.us recently estimated demand for generative AI in media entertainment at US$1.4 billion in 2023 with a projected CAGR of 26.3% through 2033. Our analysis confirms the massive potential, indicating the chance for GenAI to add 14.4 percentage points to operating profit.
Industry professionals are rapidly beginning to see the potential. Just recently, filmmaker and actor Tyler Perry halted a US$800 million expansion of his production studio after seeing a preview of OpenAI’s video generator, Sora.
The industry could soon move toward highly personalized content creation, where AI generates content based on individual viewer preferences. Although such customization at scale may present initial challenges, it could fundamentally alter how entertainment is produced and consumed.
Marketing services is another area likely to face significant change for many of the same reasons as the entertainment industry. GenAI-powered tools for marketing were among the first generation of GenAI applications available. The market is flooded with young vendors in this space and is likely to see significant consolidation in the coming months as a result. Established incumbents such as Salesforce are increasingly incorporating generative AI into their existing products as well. As in entertainment, video- and music-generation tools will offer significant cost savings for marketing services. Another similarity: the personalization that the cost and speed of generative AI enables is among the use cases offering the most transformational value, and its impact is still nascent. GenAI could even bring about a disintermediation of marketing services due to the leverage it provides to the companies they currently serve.
More significant disruption could be on the horizon as the move toward agentic systems, which act autonomously on a user’s or organization’s behalf, accelerates. One example of these signals comes in the form of a startup called Firsthand, which was founded by adtech luminaries who developed some of today’s online advertising platforms. The company is working on developing a system by which advertisers can embed AI agents into third-party content. The idea is that consumers can activate these agents directly and interact with them to ask product questions or even execute tasks. This could foreshadow an entirely new category of online advertising.
There’s little time to lose for the industries in this quadrant, making it critical for the organizations within them to build out the platforms and capabilities they’ll need to survive for the long-term. We’ll share an effective way for them to do so later in this article.
Disruptors (low ease of adoption and high level of disruption). Disruptors are industries that may find adoption more challenging due to various factors, such as regulatory hurdles or the need for significant customization. However, they recognize the critical importance of staying relevant with the prospect of major disruption on the near horizon and are thus motivated to integrate generative AI into their business models.
Take pharmaceuticals, for example. Generative AI is poised to revolutionize drug discovery and patient care by predicting molecular responses and personalizing treatments. Such innovation will significantly disrupt the market, making treatments faster to develop and more effective, potentially increasing both speed-to-market and operating profit margins dramatically.
Both new entrants and established organizations, in many cases working together, are already using generative AI drug-discovery platforms today. For example, Insilico Medicine used its Pharma.AI platform to find the right target and develop a compound, INS018_055, aimed at treating idiopathic pulmonary fibrosis—a lung condition with rising prevalence and often severe outcomes. In June 2023, the compound entered Phase II human clinical trials as the world’s first drug fully designed by generative AI (though using a generative adversarial network as opposed to a large language model). The company also has a partnership with pharma giant Sanofi to accelerate its drug development, an example of just one of dozens of such partnerships. Others include Bristol Myers Squibb and VantAI, and Merck’s deals with BenevolentAI and Exscientia. GlobalData found AI drug-discovery partnerships increased 575% from 2015 to 2020 alone.
The heavily regulated pharma industry has shown a willingness to adopt generative AI in other areas of the business as well. Pfizer, for example, recently announced its new internal GenAI-powered application, Charlie, to help its marketing employees craft digital media messages and presentations. It features a system to tag content as high-risk for further review to ensure accuracy and compliance. The company plans to integrate marketing analytics and other features going forward.
Medical technologies and private healthcare are also likely to see significant disruption, and organizations in the space are moving quickly. Numerous vendors are launching new tools aimed at helping clinicians develop clinical notes, along with plain-language versions for patients, and other applications that aid in more accurate diagnoses and interpretation of medical imaging. Incumbents are acting fast as well. Electronic healthcare records provider Epic, for example, began integrating GenAI into its tools in 2023 to assist in such areas as responses to patient queries, medical coding, and patient summaries for care team hand-offs. Research is showing generative AI could even move into a direct patient-care role in some cases, particularly in mental health, which could open the door for a much larger proportion of the population to access such care—as well as a reimagination of how care is delivered.
This fast movement in the pharma and private healthcare sectors typifies the speed with which industries in this quadrant are moving to ensure they become the disruptor as opposed to the disrupted. With changes already in motion, all organizations in this quadrant are under pressure to overcome adoption barriers.
Streamliners (high ease of adoption and lower level of disruption). Streamliners are sectors in which the adoption of generative AI is relatively straightforward and is more likely to be used primarily to enhance efficiency and productivity in the nearer term rather than to completely reinvent business models immediately. However, as evidenced by the clusters of industries in the upper left and lower right of this quadrant, some streamliners are more likely to see rapid disruption than others.
In the grocery industry, for instance, AI could be used to optimize supply chains and inventory management, reducing waste and ensuring the freshest products. Generative AI can also enhance the shopping and employee experience. Walmart, for example, is using generative AI for Text to Shop, which enables customers to ask for a product and schedule delivery through natural language via mobile texting. Other generative AI apps from the retailer provide online shopping assistance and answer customer service inquiries. And employees can now use an app called Ask Sam to quickly locate products within their store, look up prices, and find internal policy information. Use cases like these are somewhat easy to execute, and while they don’t necessarily upend the industry, they can significantly streamline operations and improve profit margins.
Another industry in the streamliner category is asset and wealth management (AWM). Many areas of businesses within AWM offer a faster track toward GenAI adoption and significant efficiencies, despite it being a highly regulated sector. These include accelerating investment research and report generation and improving customer service.
Already we’ve seen Bloomberg integrate GenAI functionality into its terminal. The new capability allows equity analysts to access summaries of the often-lengthy earnings calls and reports they must sift through, with links to deeper content for further exploration and validation. In another example, Morgan Stanley rolled out an internal tool to its financial advisors that enables them to use natural language to access information in 100,000 documents, saving time and improving their ability to serve clients. Several banks, such as Wells Fargo, have also improved automated customer service tools with generative AI.
Although efficiency plays may take center stage in AWM in the short term, development of one of the industry’s most significant disruptive services—automated, personalized investment portfolio development—is likely already underway. At least one financial services firm is reportedly working on a GenAI tool that would deliver personalized investment recommendations.
Industries in the lower right corner of the multitasker category may take a more balanced or cautious approach to AI integration, but opportunities still abound. In the power and utilities sector, for instance, adoption of AI might be slower due to regulatory and safety considerations, but over time, it could lead to more efficient energy use and reliable service provision. Organizations in this industry could implement AI to optimize energy distribution and predict maintenance needs for infrastructure.
Movement is already occurring in this space. In February 2024, the US Department of Energy published ChatGrid, a generative AI tool available for download from GitHub. It allows grid operators to create infrastructure visualizations and ask questions about the location of power plants, generation capacity, and more in natural language. Sensitive grid data wasn’t used to train the underlying model, and operators can run the tool locally where they can more safely upload their own data.
It’s important for streamliners to avoid complacency when it comes to looking for ways to leverage generative AI for more disruptive opportunities. The emergence of disruptors like Amazon in retail and Airbnb in travel and leisure in the earlier days of the internet serves as cautionary tales. And given both the breakneck pace of advancement in generative AI and the innumerable ways it can be used, we expect to see disruptors moving in at a much faster clip during this wave of technological innovation.
Multitaskers (low ease of adoption and low level of disruption). Multitaskers are industries that may see generative AI as one of many technologies that can support existing operations. The telecommunications industry, for example, already has many competing technology upgrades on its plate, such as 5G and, soon, 6G, as well as moves to equip their assets with sensors and other IOT technologies.
However, the opportunity for telecoms is sizable, with an average 13 percentage point uplift in operating margins at stake. Most of the use cases for the industry lie in sales and marketing (e.g., targeted cross-selling, customer service, and submission document generation), infrastructure management (e.g., network demand planning), and internal operations (e.g., automated invoice processing and software development).
Some telecoms have already moved on some of these opportunities. Telkomsel, for example, launched a GenAI-powered assistant, Veronika, that recommends offerings to customers and answers service inquiries. AT&T and South Africa’s MTN launched similar customer-facing chatbots. And Three UK is using generative AI to optimize its networks.
Other industries in this category, such as retail banking, may move slower in applying generative AI in some areas due to regulatory scrutiny. Hyper-personalized customer offerings could be one such use case, given heightened concerns around handling sensitive customer data. Other areas like software development, automated IT servicing, and customer service chatbots have lower barriers and high efficiency upside. Indeed, some retail banks are already using generative AI in these areas. Bank of America and NatWest are just two that have built generative AI into their customer service chatbots.
Pattern recognition paves the path to reinvention
Regardless of the potential timing of industry disruption and level of adoption challenges, we’ve identified a sector-agnostic approach to GenAI implementation that can enable organizations to capture efficiency gains while laying the groundwork for more transformative applications. It relies largely on pattern recognition.
What we mean by a pattern is the common model architecture, tooling, and design elements that enable each of GenAI’s six primary capabilities: net-new creation, augmentation, transformation, dialogue, information retrieval, and summarization (see table for a more detailed description of each). Build these elements out for one use case and they can be quite easily repurposed for other use cases that deliver the same capability. And simply thinking about applying generative AI in this way, as opposed to focusing on what specific tasks GenAI can automate, sets a flywheel in motion that both reveals and provides the foundations for implementations that ignite reinvention.
Start with the value hypothesis and potential use cases
Although we’re recommending a focus on patterns, we’re not suggesting organizations do so in lieu of identifying their most promising sources of value. As we shared in our detailed article on the generative AI flywheel (see the graphic below for a quick review), forming your value hypothesis is the first step.
This involves an assessment not too dissimilar from what organizations conducted in the past when deciding where to begin their digital transformation efforts. Find your highest sources of value, and then determine if generative AI can unlock even more value.
Although the sources of generative AI value will, of course, vary from one organization to another, our analysis indicates a starting point for exploration in each industry. Sales and marketing and core service delivery (e.g., patient care in private healthcare) are the functions in which generative AI can surface the most value for organizations in many industries, but there are nuances. For example, some disruptors are more likely to see significant value in applying generative AI in research and development. Pharmaceutical companies serve as a case in point—they could derive half of generative AI’s value potential from using it for R&D (e.g., from improving drug discovery). Among trailblazers, software companies stand to gain an equal amount of value from R&D and sales and marketing (roughly 40% from each).
Ease of adoption should then be considered to prioritize the list. The ability to execute in a responsible manner is perhaps the most important of several factors for consideration (including the level of model customization required, regulatory barriers, and cultural readiness). Skipping this step would not only leave value on the table but also lead to significant problems.
For example, a financial services company came to us after standing up a customer service chatbot, only to have to take it out of service because it was hallucinating (generating inaccurate responses). We quickly found that the organization hadn’t developed the solution with a responsible AI framework in place, which directly contributed to the problem. For example, challenging the system to see how and when it would produce biased or inaccurate responses (known as red teaming), which is one of many responsible AI practices, did not occur.
Finding the patterns to set the flywheel in motion
Once you’ve shortlisted the key use cases for your organization, it’s time to look for patterns. Which use cases leverage the same models and tooling? This is the step we find a lot of organizations skip, which can lead to a future of significant technical debt and limited value in the long run. Sometimes the single most-valuable use case may not be the right one to start with because its technical underpinnings don’t support additional use cases.
In other instances, a use case may enable this, but only if its supporting data, models, and tooling are constructed in a way that takes such flexibility into account. For example, we worked with a software and services company that came to us with the desire to create a GenAI-powered assistant that would help its agents surface policy information much faster. Such an agent could be created in a siloed manner by aggregating only policy data and applying an embedding model, vector database, and the appropriate generative model to produce summaries of key information where needed. But by looking across their longer list of use cases and identifying the predominant pattern (information retrieval), it became clear that creating a data model for a larger swath of their unstructured data would enable many additional valuable use cases in other areas of customer service and across sales and marketing. They’re now building many products using the same data model and architecture as opposed to standing these up separately using duplicative technologies and additional resource—and missing out on setting the flywheel in motion that now has them on a path to true reinvention of their operating model.
As we venture into the future, the journey for every organization converges toward a common goal: capturing the prodigious productivity gains offered by generative AI, while navigating its potential to disrupt existing paradigms. The early tremors of transformative change are already palpable in some industries, foreshadowing the widespread shifts that others will soon encounter. Our research shows that leaders are aware of the impending evolution and recognize the imperative to steer their enterprises through these massive shifts to remain competitive.
The challenge—and opportunity—lies in transcending the allure of mere automation and isolated solutions. Instead, the forward-thinking organizations will be those that discern broader patterns and leverage these insights to catalyze reinvention. This approach can prepare them to both survive the forthcoming waves of change and set the new standard for their industry in an AI-dominated future.
Subscribe now
Get s+b leadership insights direct to your inbox
Bret Greenstein is a thought leader in data and analytics and generative AI. He is a principal with PwC US.
Jennifer Kosar is in PwC US’s Digital Assurance and Transparency practice and is the US Trust and Transparency Solutions Leader. She is a partner with PwC US.
Colin Light is PwC’s EMEA and UK Leader of Strategy&, PwC’s global consulting business. He is a partner with PwC UK.
Mary Shelton Rose is PwC’s EMEA and UK Technology, Media and Telecommunications industry leader. She is a partner with PwC UK.
