AI isn’t just improving corporate productivity and efficiency. It’s poised to reshape how value is defined, delivered and captured with its ability to create content, ideas and logic, and solve problems autonomously. The result: new AI-fuelled business models are challenging long-standing business assumptions and further blurring industry boundaries, contributing to massive new value pools and growth opportunities that will emerge over the next decade and beyond.

While past tech shifts—like cloud, mobile and the internet—also enabled new business models and rapid-growth industries, AI is notable in its potential to transform the economics of product and service creation, customisation and scale with its ability to reason, to continuously learn and to use natural language to interact with humans. In many cases, the large language models at the foundation of GenAI allow companies to create additional units of output—the design of a personalised product, a customer service response or a digital offering—more efficiently, and often at near-zero marginal cost.

As a result, companies can create customer value that is less constrained by production costs, including labour. Expanding the scope of how products are produced, distributed and purchased might no longer introduce more operational complexity. And managing all types of capital—financial, physical assets and talent—may no longer be limited by data bottlenecks. Instead, companies can act on data at the speed it’s generated for adaptive decision-making.

The breakneck pace of GenAI adoption (with ChatGPT alone potentially closing in on nearly 1 billion users) and the encouraging business results offer an early indication of the technology’s promise. However, it’s still unknown how broadly and deeply these new AI-fuelled business models will take hold. Some will be heavily influenced by the pace of AI adoption, shaped by the degree of trust in the technology, supporting governance models, and emerging capabilities to verify, authenticate and validate underlying data, transactions and parties—human or AI. Nonetheless, as past technology transformations have shown, fast movers won’t just gain an edge—they’ll reset the baseline for customer experience, putting pressure on the rest of the market to follow.

To help leaders seize these new opportunities—and avoid being caught off guard—we lay out nine new business models that fall into three broad categories: scaling services, increasing product scope and access, and rapidly activating all types of capital as events unfold. We also note key business moves organisations will likely need to make and explore foundational questions leaders can ask to pivot quickly when the time is right.

Nine business models to watch

— Source: PwC

Scaling services, not size

AI is already reshaping services business models, extending expertise, advice and support without adding employees while reducing the cost per additional unit of service (per hour, session or procedure). As these services become more personalised based on customer context and history, companies will need to rethink everything from supply chains to organisational designs in order to achieve the intelligence and capacity to adapt any product to each customer’s needs.

Services as software. Companies that sell physical products can now offer real-time, contextual AI services alongside them—across thousands of product variations. Doing so equips organisations to turn one-time sales into ongoing customer value while reducing the cost per service hour. A consumer packaged goods company, for instance, might offer an AI nutrition assistant that doesn’t just cite product nutritional data, but answers real-world questions like ‘Which snack is a good option if I’m avoiding gluten and need to lower my blood sugar?’ or ‘Which protein bars are best before my Tuesday weight training and Thursday cardio classes?’ These services could be monetised through micro-payments—priced per use, query or output—or embedded in the product experience to boost differentiation and perceived value.

Companies that sell physical products can now offer real-time, contextual AI services alongside them—across thousands of product variations. Doing so equips organisations to turn one-time sales into ongoing customer value while reducing the cost per service hour. A consumer packaged goods company, for instance, might offer an AI nutrition assistant that doesn’t just cite product nutritional data, but answers real-world questions like ‘Which snack is a good option if I’m avoiding gluten and need to lower my blood sugar?’ or ‘Which protein bars are best before my Tuesday weight training and Thursday cardio classes?’ These services could be monetised through micro-payments—priced per use, query or output—or embedded in the product experience to boost differentiation and perceived value. The new agentic AI advisors. Agentic AI advances the capabilities of traditional AI agents, creating systems that don’t just respond to inputs, but proactively solve problems, coordinate tasks and continuously learn through their work. Companies that offer industry-focused advisory services, such as financial, wealth management, health and wellness, and legal services, can use agentic AI to create autonomous AI team members (we call them agentic AI advisors) that can guide human advisors and customers at a fraction of the cost of human staff. Today’s task-specific B2C AI agents, such as AI-powered budgeting apps and AI personal health assistants, represent the first wave of offerings in this area. Expect organisations to create even more customised packages using humans and agentic AI capabilities to span multiple areas (for example, combining health and financial planning) and linking advice directly to actions, such as executing a financial transaction based on the guidance. These services can be scaled to benefit hundreds of thousands of customers on a subscription-based, fee-for-service or per-session basis.

Agentic AI advances the capabilities of traditional AI agents, creating systems that don’t just respond to inputs, but proactively solve problems, coordinate tasks and continuously learn through their work. Companies that offer industry-focused advisory services, such as financial, wealth management, health and wellness, and legal services, can use agentic AI to create autonomous AI team members (we call them agentic AI advisors) that can guide human advisors and customers at a fraction of the cost of human staff. Today’s task-specific B2C AI agents, such as AI-powered budgeting apps and AI personal health assistants, represent the first wave of offerings in this area. Expect organisations to create even more customised packages using humans and agentic AI capabilities to span multiple areas (for example, combining health and financial planning) and linking advice directly to actions, such as executing a financial transaction based on the guidance. These services can be scaled to benefit hundreds of thousands of customers on a subscription-based, fee-for-service or per-session basis. Robotic service providers. Imagine the next generation of robotic vacuum cleaners able to understand and fulfil a request like ‘clean up the spot in the living room where we tracked in dirt,’ without any further direction. Robotic lawn mowers that could autonomously adjust their schedules based on weather reports and yard usage. Robotic nursing assistants that could alert hospital staff when a patient is confused. These are a few of the advanced AI-powered aides that industries, including facilities maintenance and construction, could create to perform labour-intensive, dangerous or precise manual tasks. These robotic aides represent more than just a product upgrade. Because of their ability to learn user preferences and dynamically respond to changes in the environment, they can be hired out to provide ongoing services through product leases or subscriptions (charged per task, per engagement or per hour), or generate revenue from the data they collect.

Increasing product scope and access without complexity

Emerging business models in this category use AI’s capability to deliver more personalised products—both what you offer and how you fulfil it—with fewer operational trade-offs. Each model in this category could deliver significant value on its own. Combined, they could create a supply network that adapts in real time and deliver hyper-personalised products to customers everywhere.

Mass customisers. Some personalisation at scale is already here, with manufacturers providing customers with dozens of choices, of colours, patterns, finishes, materials and other options. The mass customisers of the future, however, will personalise products at incredible speed and with the kind of customer relevance that no menu of options can offer today. Think personalised vitamin packs based on blood panels, or modular furniture manufactured based on your room dimensions, lifestyle and taste—at your doorstep in days, not weeks or months. AI will be able to dynamically tailor designs, materials and fulfilment for each customer, and orchestrate production and distribution across partners or internal systems, without increasing cost or delivery times and without sacrificing margin.

Some personalisation at scale is already here, with manufacturers providing customers with dozens of choices, of colours, patterns, finishes, materials and other options. The mass customisers of the future, however, will personalise products at incredible speed and with the kind of customer relevance that no menu of options can offer today. Think personalised vitamin packs based on blood panels, or modular furniture manufactured based on your room dimensions, lifestyle and taste—at your doorstep in days, not weeks or months. AI will be able to dynamically tailor designs, materials and fulfilment for each customer, and orchestrate production and distribution across partners or internal systems, without increasing cost or delivery times and without sacrificing margin. Reverse auction marketplaces. Imagine asking an AI shopping agent to find a phone with specific features within a budget—and receiving competitive offers. In this model, consumers broadcast what they want and are willing to pay, and sellers bid to win the sale—flipping the script on traditional e-commerce dynamics. We’re already seeing early signs of this shift in models like ‘name-your-price’ insurance, legal platforms that match lawyers to fixed-fee cases, and dynamic ticketing systems that adjust prices based on demand. The next wave of agents won’t be restricted to a particular category; instead, they will search and transact this way across all types of categories, reducing transaction costs and friction during retail purchase without the need for more staff.

Imagine asking an AI shopping agent to find a phone with specific features within a budget—and receiving competitive offers. In this model, consumers broadcast what they want and are willing to pay, and sellers bid to win the sale—flipping the script on traditional e-commerce dynamics. We’re already seeing early signs of this shift in models like ‘name-your-price’ insurance, legal platforms that match lawyers to fixed-fee cases, and dynamic ticketing systems that adjust prices based on demand. The next wave of agents won’t be restricted to a particular category; instead, they will search and transact this way across all types of categories, reducing transaction costs and friction during retail purchase without the need for more staff. Autonomous delivery anywhere. This model expands how quickly—and where—goods and services can be delivered using increasingly autonomous fleets of cars, trucks, drones, aircraft and ships. Imagine the possibilities for a retailer or manufacturer when a cargo truck can run 24/7, drones can make last-mile deliveries, and self-driving delivery vehicles are in every town. Same-day, even same-hour, delivery becomes accessible for more businesses and in more locations, without traditional labour constraints (and while reducing traffic congestion and energy usage). More dynamic customer experiences, such as being able to alter an order mid-delivery, will likely emerge. Companies can charge usage fees (such as ride fares, delivery fees or leasing autonomous machines) or higher prices for autonomous-capable integrated distribution and product models. This business model has the longest on-ramp because AI-powered vehicles must reach a high level of safety and efficiency, which they can only do by encountering a lot of situations and learning from the data.

Managing capital with greater precision and without the data drag

The previous two categories focus on how companies will likely create and deliver products and services to expand their offerings, but this last category of business models addresses how companies will optimise capital decisions related to financial assets, physical assets and talent. These business models tap into AI’s ability to glean contextual cues from huge volumes of data, simulate outcomes, and become smarter and more efficient with every user, transaction and interaction—enabling companies to monetise high-frequency activity that has been too complex to manage. Because these business models are rooted in broad collaboration across industries, regions and businesses, they will be more dependent on trust and trust solutions—with shared standards and automated, scalable mechanisms for authenticating the participants and the data at every layer of the exchange.

Precision capital allocation-as-service. Quantitative hedge funds already demonstrate what reliable AI-enhanced algorithmic trading looks like by optimising portfolios at enormous scale. This business model taps into AI to expand how and where financing organisations can deploy capital to match financing and investment needs in real time across a portfolio of projects. It can empower, for example, a business to coordinate hundreds of thousands of loans in an AI-powered lending marketplace in real time, adjusting rates based on individual risk or dynamically personalising repayment terms based on an individual’s or company’s monthly income. This type of offering can be applied across corporate investments, loans, venture funding, insurance contracts, and even public budgeting and talent financing.

Quantitative hedge funds already demonstrate what reliable AI-enhanced algorithmic trading looks like by optimising portfolios at enormous scale. This business model taps into AI to expand how and where financing organisations can deploy capital to match financing and investment needs in real time across a portfolio of projects. It can empower, for example, a business to coordinate hundreds of thousands of loans in an AI-powered lending marketplace in real time, adjusting rates based on individual risk or dynamically personalising repayment terms based on an individual’s or company’s monthly income. This type of offering can be applied across corporate investments, loans, venture funding, insurance contracts, and even public budgeting and talent financing. Dynamic asset-monitoring utilities. Predictive maintenance isn’t new. But AI’s ability to efficiently make sense of second-by-second complex sensor and IOT data, extract insights from broad data sets (including historical data), and identify cause and effect will support the delivery of new asset-monitoring utilities that help owners prevent adverse outcomes for any asset (e.g., a factory machine, an oil pipeline or an office building) that can break, overheat, catch fire, leak or otherwise pose risk. These asset-monitoring utilities could, for example, alert a property manager that a water heater is experiencing increased power usage and unusual vibration and has a 90% probability of causing a flood event within 48 hours. Companies offering these services can charge asset owners or insurers a per-site or per-asset service fee or implement analytics-as-a-service contracts, helping their customers to improve maintenance schedules and reduce unplanned downtime, safety incidents and, potentially, insurance premiums, while aligning costs with actual consumption.

Predictive maintenance isn’t new. But AI’s ability to efficiently make sense of second-by-second complex sensor and IOT data, extract insights from broad data sets (including historical data), and identify cause and effect will support the delivery of new asset-monitoring utilities that help owners prevent adverse outcomes for any asset (e.g., a factory machine, an oil pipeline or an office building) that can break, overheat, catch fire, leak or otherwise pose risk. These asset-monitoring utilities could, for example, alert a property manager that a water heater is experiencing increased power usage and unusual vibration and has a 90% probability of causing a flood event within 48 hours. Companies offering these services can charge asset owners or insurers a per-site or per-asset service fee or implement analytics-as-a-service contracts, helping their customers to improve maintenance schedules and reduce unplanned downtime, safety incidents and, potentially, insurance premiums, while aligning costs with actual consumption. Talent on tap. This model anticipates a world where almost any service—from skilled consulting to everyday errands—can be obtained on a pay-per-use basis. It goes beyond current gig platforms, which are often limited to one type of service, like ride-hailing or freelancing. Instead, these platforms, which will likely be based on commission fees, use AI to handle the heavy logistics for every type of job a company may need, drastically reducing friction for both providers and consumers. AI’s contributions include finding available providers (including freelancers, contractors, agencies and firms) with the desired skill, scheduling them, setting a price and handling payments on demand, and ensuring quality through ratings or monitoring.

How to prepare for this future

The business models outlined above aren’t linear or incremental extensions of today’s approaches. They would fundamentally change the economics of customer value creation—and businesses that embrace them could reshape the market dramatically. A good way to understand what this could mean for your business is to start with these four questions.

How will our fiercest competitor use AI to beat us? Imagine a start-up with deep pockets and no legacy workflows, systems or technologies promising to reinvent your industry with AI. Conducting structured war-game scenarios or red-versus-blue team exercises with a mix of leaders from strategy, marketing, sales, finance, operations, supply chain and product development can help ensure organisations think boldly enough about the potential disruption, including where its strategy, investments, and workforce must evolve and what impact AI can have on revenue, margin, and the workforce. It can also result in actionable initiatives, such as identifying use cases to test new opportunities and surface risks. Our experience suggests that companies that approach business model reinvention iteratively in this way are most successful.Once the strategy is clear, companies can then benchmark competitor investments and decide whether to lead, follow quickly or adopt cautiously based on the business potential, associated risk and required investment. They can also assess how much to realistically invest given projected AI-driven growth and margin improvement, investor return expectations and flexibility to reinvest capital in long-term value creation. In some cases, this effort could lead companies to raise capital or change their posture for reinvesting in the business. How will our customer experience change with AI assistants in the mix? Mapping your customer journey as if both sides—your company and the customer—are interacting through intelligent agents can help identify the kinds of experiences that will attract both customers and their AI assistants, and which could break under new expectations. If a customer’s AI assistant can’t access order status, warranty info and return policies for multiple products simultaneously due to your company’s security policies, your products may not end up on the AI’s list of recommended products to review.For every touchpoint, assess the marginal cost to deliver and the potential revenue it could generate. This can help define what a profitable customer experience looks like in an AI-powered market—and clarify which interactions are essential. The same analysis can also surface pain points driving retention issues today, such as inadequate service capacity, unclear pricing, product issues or low-value features. How will production and fulfilment run in a hyper-personalised world? A customer journey map can also help your organisation prioritise the extent to which personalisation options drive value—so you customise only what matters most. For instance, service providers could use satisfaction scores and Q&A data from across the customer journey to identify ways AI workflows could dynamically route service representatives based on the performance targets or specifications for installed equipment at each customer site. The impact: increasing customer satisfaction, optimising the value delivered for each available service hour and, potentially, enabling premium pricing.Once the customer journey is understood, companies can turn inwards to analyse the full value chain—step by step and cost by cost for each product and service, identifying barriers to customisation, such as rigid manufacturing systems and product architectures, siloed sales and production processes, or supplier limitations. The result is a practical blueprint that prioritises the production workflows and technology to be reinvented and those that require only basic efficiency improvements for delivering AI-driven personalisation. What will your workforce look like with human and AI agents side by side? In an interview, OpenAI’s Sam Altman described a standing wager with other leaders about when the first company will reach a US$1 billion valuation with just one employee—supported by AI agents. That might sound extreme, but the message is clear: AI agents can take on specialised tasks at scale, augmenting human roles throughout the organisation. A useful exercise is to assess what kinds of AI agents could complement each major role in your organisation, which new roles might emerge and how those roles could add value within these new business models. The findings can help leaders understand how their workforce is likely to evolve and map the capabilities needed for successfully deploying agentic AI capabilities, including infrastructure investments to protect data privacy and data security, as AI agents enter the mix, and governance and process changes that ensure the right level of human oversight. Regardless of whether AI delivers a sudden leap in performance or more gradual changes, companies can benefit from upskilling their workforce to understand how AI agents can help increase their productivity and reduce organisational sludge.

Although the list of increasingly powerful large language models continues to grow and hundreds of millions of people use them in their daily lives and at work for a variety of tasks, the readiness to use GenAI to reshape how value is created, delivered and captured varies dramatically across industries and companies. So too does the level of trust. Lower-risk innovations, such as an AI-powered robot vacuum or AI product advisors, are poised to leapfrog first as they may be able to grow in most environments, whereas higher-risk autonomous decision-making systems will likely require governments, industries and society to align on responsible use of AI worldwide. Evaluating what your biggest competitor might do, how your business may evolve and what capability investments are needed can help your organisation not only contemplate (and prepare for) new AI-enabled business models’ effect on your future viability but also surface opportunities to improve how you compete today.

View source