Business Models in the AI Era

Given the speed with which AI is developing and the opportunities it offers to reinvent companies, we are sharing our predictions of how business models—how companies make money—will evolve over the next five-plus years. We started this work by analyzing MIT CISR studies from 2013 to 2025 (with data from 2,378 companies) on how business models have evolved.[foot]MIT CISR collected business model data from companies in 2013 (101 companies in one survey, 93 in another), 2019 (1,311 companies), 2022 (721 companies), and 2025 (152 companies), each time using the same questions and placing companies on MIT CISR’s Digital Business Models framework, described in Peter Weill and Stephanie L. Woerner, What’s Your Digital Business Model?: Six Questions To Help You Build the Next-Generation Enterprise (Harvard Business Review Press, 2018). The methodology places each company into one business model, but companies can operate in multiple business models, as our case vignettes and interviews show. In 2025, to contemplate business models five years out and beyond, we convened a group of four researchers (including a former CIO) to imagine how the key dimensions differentiating business models will evolve with the rise of advanced AI (e.g., generative AI, agentic AI, robotic systems) and formulate a new framework. We then tested the framework by presenting and discussing it in a series of senior executive meetings.[/foot] For example, the percentage of companies in our research that lead or participate in a digital ecosystem has increased from 30 percent in 2013 to 81 percent in 2025, with commensurate increases in revenue growth and net profit. This dramatic shift illustrates how in just twelve years digital technologies have revamped business models. We expect the next decade will see an even more dramatic change as companies develop AI-driven, real-time business models.

In this briefing we describe how business models have developed over the last twelve years and share our business model framework for the AI era. We then illustrate how companies are navigating these changes with a case study of One New Zealand.

Business Model Performance 2013–2025

Starting in 2013, using a digital business model framework we developed, we categorized the companies from our research into four digital business models[foot]See MIT CISR’s original Digital Business Models framework on the MIT CISR website at https://cisr.mit.edu/publication/whats-your-digital-business-model#dbm.[/foot]:

• Supplier: Sells via other companies (e.g., manufacturer, insurance via an agent)
• Omnichannel: Combines digital and physical channels (e.g., retailer, bank)
• Modular Producer: Provides a plug-and-play product or service (e.g., payments provider)
• Ecosystem Driver: Offers a go-to destination in a customer domain[foot]MIT CISR defines a customer domain as a customer’s end-to-end need in an area such as home, mobility, energy, education, corporate services, or secure supply chain; see P. Weill and S. L Woerner, “Top-Performing Companies Focus on Customer Domains,” MIT CISR Research Briefing, Vol. XXIII, No. 9, September 2023, https://cisr.mit.edu/publication/2023_0901_DomainOriented_WeillWoerner.[/foot] (e.g., buying and owning a house, managing energy use), connecting customers with providers

We collected data on companies’ prevalent model in 2013, 2019, 2022, and 2025, which showed a gradual shift to Modular Producer and Ecosystem Driver. From 2013 to 2025, prevalence of Supplier decreased from 46 percent of companies to 15 percent, Omnichannel dropped from 24 to 4 percent, Modular Producer grew from 18 to 23 percent, and Ecosystem Driver increased from 12 to 58 percent. Companies that lead ecosystems can deliver integrated solutions and seamless interactions, which customers have increasingly expected. The success of this approach is evident, as in our research Ecosystem Driver was the only digital business model in 2025 with above industry average revenue growth (by six percentage points).

Business Models: 2025 and Beyond

With the explosion of AI technologies—including machine learning and generative, agentic, and robotic AI—we anticipate that business models will evolve to become increasingly outcome oriented and enabled by autonomous AI. Therefore, we’ve created a new version of our business model framework for the AI era—see figure 1.

Figure 1: Business Models in the AI Era

Source: In 2025, to contemplate business models five years out and beyond, MIT CISR convened a group of four researchers (including a former CIO) to imagine how the key dimensions differentiating business models will evolve with the rise of advanced AI (e.g., generative AI, agentic AI, robotic systems) and formulate a new version of our Digital Business Models framework for the AI era (above). We then tested the framework by presenting and discussing it in a series of senior executive meetings.

The vertical axis of the framework, previously knowledge of your end customer, has evolved to action on behalf of customers. Companies either assist customers, playing a supportive role in helping them reach outcomes, or represent them, autonomously achieving outcomes for them within guardrails.

The horizontal axis has refocused from business design (value chain or ecosystem) to business execution, with companies taking either a structured or adaptive approach. The structured approach starts with a specified business process and builds toward outcomes, with AI executing predefined flows of activity. Employees review and approve all decisions before execution—i.e., human in the loop; and set strategic goals, ensure incentives are aligned, define constraints and guardrails, and monitor outcomes—i.e., human at the helm. The adaptive approach specifies an outcome and empowers advanced AI to construct an effective process to achieve it, with humans at the helm.

Combining these two axes produces four business models that rely on increasing levels of automation and autonomous AI: Existing+, Customer Proxy, Modular Curator, and Orchestrator.

Existing+: In this model, the company augments its existing business model with AI. It assists customers in achieving outcomes by delivering products and services through AI-enhanced established processes—for example, when a financial services company uses AI to enhance its traditional advisory process by analyzing customer data and providing more personalized investment recommendations.

Customer Proxy: Here, the company represents customers to achieve outcomes for them using predefined processes, with AI supporting execution. An example is when a financial services company uses AI to automatically manage a customer’s portfolio within predefined parameters.

Modular Curator: With this model, the company assists the customer in achieving outcomes by using AI to adaptively assemble reusable, combinable modules, including those from other companies, into tailored service bundles that meet the customer’s goals. For example, a financial services company might use AI to assemble and recommend a personalized financial bundle that combines investment products, insurance, and credit solutions from multiple providers based on the customer’s goals.

Orchestrator: In this model, the company represents the customer to autonomously achieve outcomes on their behalf through adaptive, AI-mediated collaboration across an ecosystem of products and services. For example, a financial services company might offer a fully managed wealth solution where AI continuously optimizes the customer’s portfolio to meet long-term goals without requiring input.

An emerging example based on the Orchestrator model is Amazon’s beta AI-enabled “Buy for Me” feature that helps customers find and buy products from other brands’ sites for products Amazon doesn’t sell. A customer taps the Buy for Me button on the product page to request that Amazon make the purchase on their behalf from the brand retailer’s website.[foot]“Amazon’s new ‘Buy for Me’ feature helps customers find and buy products from other brands’ sites,” Amazon News, accessed September 14, 2025, https://www.aboutamazon.com/news/retail/amazon-shopping-app-buy-for-me-brands.[/foot]

One NZ is an example of a company that has begun exploring these new AI-enabled business models.

AI-Driven Business Model Innovation at One NZ

Telecommunications provider One New Zealand Group Ltd (One NZ) is making a strategic shift to move from a traditional telco to an AI-driven company. In 2024, One NZ deployed fifteen AI use cases that had reached a margin hurdle. By the end of 2025, the company aims to deploy fifty AI solutions (thirty of them have already launched or are underway) to target customer action, network optimization, and process transformation.

Most of the company’s initiatives to date reside in the Existing+ business model quadrant, with One NZ assisting customers in achieving outcomes through process-focused AI support. For example, knowledge agents assist customers with complex products, enabling over 60 percent resolution of frequently asked questions and other queries from consumer prepaid customers—and 40 percent of such queries from enterprise customers (still in beta). In another use case, marketing copywriting agents, such as an SEO optimizer and a non-visual content agent, accelerate campaign creation (e.g., by creating customer audience segments 60 percent faster) and boost speed to market.

One NZ is moving toward the Customer Proxy model, where AI agents represent customers by taking action with the customer’s authorization. Service agents already handle tasks such as upgrading plans, raising disconnection requests, and initiating support case tickets. These AI agents still have humans in the loop and humans at the helm to ensure responsible use of AI across the process. Looking ahead, these agents will also predict and resolve issues proactively.

One NZ is gearing up for adaptive business models by evolving their agents in network optimization and marketing and adding new ones. In alignment with the Modular Curator model, the company recently used a set of task-based AI agents during a major weather event to verify power failures and cell status, understand required battery capacity, forecast demand, estimate time until generator support would be needed, and recommend optimal actions to decision-makers. AI enables smarter, faster decisions and actions, allowing the company to enhance customer experience by completing in minutes activities that would previously have taken hours and permitting a better understanding of the root causes of issues.

In the future, those agents will autonomously optimize responses to power outages, positioning One NZ in the Orchestrator model. In the company’s marketing efforts, the vision is for AI agents to replace manual processes with autonomous action, first creating personalized customer campaigns, then ultimately deploying and adapting campaigns based on customer behavior. Marketing teams would specify goals, constraints, and guardrails, and then monitor outcomes.

What’s Next for Your Company

With rapid growth of AI technologies, it’s time to think about how your company will adapt to thrive in the AI era. Most companies face pressures on margins and growth, making it critical to identify opportunities to create new value. Consider these questions:

1. Looking at figure 1, where can your company create the most value in the next three years?
2. Is there an advanced example of an AI-enabled business model or initiative in your company you can learn from and scale?
3. What new capabilities must you build to succeed?

In our work on digital business transformation, we have found that creating a common language is essential for progress. Do your colleagues share a clear understanding of the threats and opportunities AI presents? We recommend sharing this briefing with a group of your visionary senior leaders and facilitating a conversation of where your company stands today—and where it needs to go next.

© 2025 MIT Center for Information Systems Research, Weill, Sebastian, Woerner, and Benedict. MIT CISR Research Briefings are published monthly to update the center’s member organizations on current research projects.