It has been less than a decade since Salesforce paved the way for the now de facto enterprise software as a service (SaaS) business model, which advocates for subscription-based pricing. Key to this model is the premise that the productivity of an enterprise is proportional to the size of the workforce and therefore software can be distributed based on headcount. This is all poised to fundamentally change in the new era of enterprise artificial intelligence (AI) applications.
AI is arguably the most sought after new capability in the enterprise. In the coming years, the effects of AI will be propagated across sales, customer support, marketing and virtually all important operational aspects of running a business. With over 40% of enterprise companies now piloting these capabilities, the traditional enterprise SaaS vendors are adapting to these needs and deploying new offerings – Salesforce with Einstein, Oracle with Intelligent Bots, Microsoft with Azure Cognitive Services. Each of these companies has built a distribution model around seat-based pricing, where revenues increase as more people in the workforce use the vendor’s software. But what happens when AI product capabilities start to increase the output of employees, reducing the need for more seats? This is the dilemma that many SaaS vendors are beginning to deal with.
AI brings efficiency to digital technologies. Whether it be emailing a customer, analyzing a marketing campaign, or segmenting a sales pipeline, people can do more when augmented with AI. In industries like customer support, we’re seeing call center agents able to handle 30% more conversations with assisted AI. While this is generally good news for businesses, the vendors that service these companies will find a new challenge ahead in the sales department. Specifically, the Chief Revenue Officer will need to innovate on the business model to keep up with the innovation happening with the product. Said bluntly, why would a Salesforce account representative promote Einstein if it’s going to reduce the number of seats on an account renewal? This is a fundamental misalignment that will need to change.
A better approach would be to have the sales team aligned with the product offering. With AI at the core of new SaaS products, a natural alternative to seat-based pricing is to price based on performance. This can be done in a number of ways. One approach would be to transactionalize the work so that software is sold in units of measurable outcomes. LivePerson (NYSE:LPSN) provides a good example of this in their new LiveEngage platform. Whereas LivePerson has traditionally sold seats of their software to customer support departments, with LiveEngage, companies now pay a fee for each conversation that the company has with a customer. This allows LivePerson to align with AI partners that increase the productivity of support agents who would otherwise conflict by reducing the need to purchase more seats.
Transaction-based pricing models are also interesting because they align with longer term trends in hyper-specialization, where work is increasingly being discretized into microtasks for the workforce. If we consider this change in business model along a spectrum, we can observe that this is merely a continuation toward taylorism. Where SaaS moved the enterprise software industry away from site licenses and provided more granularity into procurement, transaction-based offerings will go deeper. Ultimately, this will be advantageous to all parties. Having measurable alignment between software tools and the workforce will drive utilization because companies will be able to better quantify the return on investment (ROI) of their software deployments and invest accordingly.
The next wave of enterprise SaaS vendors will begin offering new pricing models that align with their AI products. We can expect to see a shift away from per seat models toward models that promote ROI and utilization. This is one of the many changes we can expect to see as AI matures across industries. For enterprise SaaS vendors, the price game has been reset. Off to the races.