AI and new business models

History has shown that new technology enables new business models, and AI is no exception.

Time and again, history has shown that technological advances mean that existing structures break down and new ones emerge. Often, this means disrupting business models and entire industries, and AI - especially generative AI - is doing just that.

Beyond the productivity gains of working faster and with more quality when assisted by AI, generative AI is challenging existing business models and enabling new ones.

I'll tell you more about that here via:

1. Historical examples of how new technology has led to new business models that matched the real needs of users.
2. Examples of how AI disrupts industries today and enables new business models that better match why customers use products and services.
3. Inspiration on assessing whether your industry or business model is affected by AI.

AI is already creating new business models and phasing out old and outdated ones. But first, let's look at some examples where new technology has resulted in new business models in the past.

File sharing and music

At the turn of the millennium, file-sharing service Napster proved that people wanted easy access to the world's music so much that they were willing to break the law and download music for free and illegally via Napster.

Record labels and artists feared that we would no longer be willing to pay for music, which posed a fundamental threat to their primary source of income at the time: album sales.

However, it turned out that what we did not like was buying physical copies of albums as CDs. Music streaming services like Spotify have proven that we prefer to pay for easy access to the world's music rather than download it illegally.

According to Spotify founder Daniel Ek, as of July 24, 2024, it had more than 246 million paying users.

While we are willing to pay for music, new technology has shown the music industry that we prefer a business model and delivery model where we pay a subscription for easy access to the world's music via the Internet rather than a model where we pay for physical copies of individual albums (CDs).

Cloud and Software as a Service

With the introduction of cloud services such as Amazon Web Services (AWS) in the early 2000s, companies no longer needed to host their software on their servers. Instead, they could outsource hosting to third-party services such as AWS, which provides server space for a fee.

At the same time, internet speeds were so fast that you could use programs via the browser on your computer instead of buying physical copies. The combination of faster internet speeds and cloud services gave rise to the business model that has dominated the software market for the last 15-20 years. Instead of buying a CD-ROM with the software, you can now subscribe to it and access it via the Internet; this is now known as Software as a Service (SaaS).

The new business model led to an explosion of new companies, such as Salesforce (founded in 1999) and Zendesk (founded in 2007), that grew extremely fast via the SaaS model, where their customers pay a subscription per employee who has access to Salesforce or Zendesk products.

The business model's extremely positive reception in the market showed that companies preferred to pay for access to the software and receive regular updates rather than buying physical copies. Thus, the SaaS business model was born.

Since software vendors can sell infinite subscriptions to the same software with almost no variable costs and distribute it over the Internet, SaaS has become an extremely attractive business model for startups and venture capital.

The attractive business model and the influx of venture capital have made SaaS one of the fastest-growing and largest tech industries. Still, the disruptive business model is being disrupted now.

It's one thing that the SaaS model showed that companies prefer to buy software on subscription, but what does it mean for the business model when the software they purchase access to is no longer just a tool but now starts doing the work for them? That's what AI does, and it fundamentally changes how businesses will pay for software.

AI disrupts SaaS

Companies with the SaaS business model grow by selling subscriptions, a price per employee accessing the software. This model worked well when software was a tool that employees used to do tasks themselves. At the same time, it's a significant growth engine because as the SaaS provider's customers grow, they will naturally need more subscriptions, which means more revenue for the SaaS provider.

This user-based business model has worked for 20 years, but now we see AI challenging it. With generative AI, the product shifts from just a tool you use to do your work yourself to doing all or part of the work for you.

In customer support, where Zendesk is the market leader, companies increasingly use AI, such as AI chatbots, to automate answering customer inquiries instead of manually answering them.

One of the economic benefits companies see is that they can handle more requests with fewer support staff. This poses a fundamental threat to the traditional SaaS business model because when the software is doing the job, such as answering customer queries, companies need fewer employees to perform these tasks and, thus, fewer subscriptions with software vendors.

Customer support software vendors, such as Zendesk, are seeing revenue per customer drop, a fundamental threat to their business. But if you look closely, you could argue that it's a fundamental threat to the SaaS business model but not to the service being sold and the task the product solves.

Like the Napster/Spotify example, you need to focus on what customers really want and then develop your business around that—even when new technology disrupts the status quo.

It turns out that companies, first and foremost, want to get a given task done satisfactorily with minimal time and resources. Suppose the software you pay for can suddenly do that autonomously. For Zendesk, this means that instead of customers paying for access to the software to solve a given task (price per user), they may be willing to pay when the software performs the task itself (price per result).

This new business model, Result as a Service (or RaaS), evolves the SaaS business model enabled by AI.

Therefore, companies in the customer support industry, such as Zendesk, Intercom, and others, are rapidly experimenting with a results-based business model (RaaS model) in addition to the traditional SaaS model.

We see this in the form of hybrid SaaS and RaaS models, where companies pay for a mix of access to the software and the tasks it solves. Below is an example of Intercom's hybrid SaaS+RaaS pricing model and an explanation of how an outcome delivered by their AI is defined:

As more businesses see the benefits of paying only for results - not just access - the RaaS model will become more widespread.

The RaaS model offers an attractive future for businesses. In this model, companies pay only for the results achieved instead of access to the software they use to perform the tasks. New technology—AI—makes this future possible.

AI disrupts online media

Building business models around interesting content is at the heart of the commercial Internet - Web 2.0.

The business model for community platforms/social media like Reddit and news media like the Financial Times (FT) is to attract users to their platforms with interesting content and monetize through advertising or subscriptions.

This business model is based on the premise that users can only access specific content on Reddit's or FT's websites and apps and prefer to visit Reddit or Financial Times instead of other online media.

With generative AI products like ChatGPT, this business model is now under pressure as users of ChatGPT and similar products rely on them for the same knowledge that they would previously turn to Reddit and the Financial Times for. The providers of ChatGPT (OpenAI) and similar products have trained their language models on all kinds of web content, including content from media outlets like Reddit and the FT.

The fact that we can now get some of the knowledge via ChatGPT that we would previously visit Reddit or FT for will mean less traffic to these media. Fewer visitors mean fewer free users to serve ads to and fewer paying users to charge for access to content. This results in less revenue and poses a threat to media with a visitor-based business model.

But again, you have to separate the business model from the actual need you solve for your customers.

In the case of Reddit and Financial Times, the business model may be based on you visiting their websites, but you come there to consume the content published there.

It is the content that matters to you, not where you access it from. Companies like Reddit and FT have recognized this.

At the same time, generative AI providers like OpenAI know that quality data is one of the keys to delivering better and better models, and they are willing to pay for it.

A new business model emerges out of necessity and is enabled by new technology; Reddit and Financial Times sell their unique content to OpenAI and similar companies. A business model based on the core of their value proposition: continuous publishing of quality content, but targeting a new customer segment, such as companies like OpenAI that are willing to pay for quality data because it improves their product.

In connection with its IPO in March 2024, Reddit announced agreements with OpenAI and Alphabet (the company behind Google), among others, for access to content from Reddit:

In April 2024, the Financial Times announced an agreement with OpenAI that gives OpenAI access to content published by the Financial Times:

Reddit's and FT's new business models match the real needs of the original users: access to quality content. Thus, both companies maintain the business model based on users visiting their platforms while monetizing when users use products like ChatGPT.

Long-term consequences - the bull and bear cases

The question is whether the two business models will complement or cannibalize each other in the long run.

A possible negative consequence is that over time, users will visit Reddit and FT less and less and only do so when the user experience offers something different and better than just getting the answer from ChatGPT.

One possible positive effect is that Reddit's and Financial Times' unique content is more widely available, with a new underlying business model that matches users' real needs.

Reddit's and FT's new way of monetizing content is commercially attractive for all online media with differentiated and unique content, licensing content to AI vendors who need continuous access to quality data for training models and other uses in their products.

Customer and technology understanding is a prerequisite

One thing to consider is whether your business model is based on something structural that new technology is challenging or whether it is based on real customer needs.

• Record labels' business model was structurally based on selling physical products, such as CDs. This was challenged when the same music could be downloaded illegally online. However, it turned out that the real need was not free music but easy access to music worldwide. We would instead subscribe to easy access rather than pay for physical copies or download the music illegally. New technology provided that insight, and the music industry adapted its business models accordingly.
• Zendesk's original business model was based on access to software, but customers' actual need was to solve specific tasks. This became clear when Zendesk's product—via AI—could perform these tasks.
• For Financial Times, the business model was based solely on the number of visitors, but their customers' real need is to consume FT content rather than to visit the website. This is demonstrated by the new business model where OpenAI and the like buy access to data that can now be consumed at the FT and, e.g., ChatGPT.

Knowing your customers' real needs and how to meet them allows you to evolve your business model as new technology emerges. This adaptability is a crucial competence when it comes to new technology. History shows this, and we see it today with AI.

Understand so you can (re)act

Businesses must respond when new technology disrupts existing business models and enables new ones. They must both embrace the change it represents and seek to understand how new technology—like AI—supports the real needs they solve for their customers.

The best way to do this is to get your hands on the technology and gain experience using it in practice.

As FT Group CEO John Ridding says about the agreement with OpenAI: