Pricing Generative AI Features for SaaS Products: Finding the Right Path
By Go-To-Market Operations
The growing trend of integrating generative AI (GenAI) into SaaS businesses has unlocked what we believe are exciting opportunities for software providers to offer more intelligent, interactive, and efficient solutions to their customers. However, one critical challenge has emerged: how companies can effectively price GenAI features within their existing pricing structures.
SaaS companies have traditionally followed a subscription-based pricing model. This model builds stable revenue streams and fosters long-term customer relationships. In the last few years, there has been a growing emphasis on hybrid pricing, a blend between the traditional subscription and an element of usage-based or value driver-based pricing. With the acceleration of GenAI in the last 12 months, this movement towards hybrid pricing is perhaps accelerated. Having an open tab for Open.AI – sending prompts to ChatGPT as needed – can incur new costs and potentially erode profit margins due to OpenAI’s pricing model, which is based on usage. While Open AI dropped their per-token pricing significantly about a year ago, hybrid pricing has a lot of merits, and we want to explore the topic more in depth. As such, it may become necessary for a traditional subscription-only solution to embrace more of a usage-based model, at least in part.
Companies should consider what a potential hybrid pricing model looks like for their business. These models can benefit companies where scaling COGS is a challenge or a variable cost product or feature is being deployed (e.g., GenAI). It’s an important exercise to regularly re-evaluate customer value drivers and identify where, and how much, they can charge for the platform subscription (i.e., access to the software) and where they can charge for driver-based elements. Like hosting costs when migrating to the cloud, GenAI usage volumes may be insignificant for some businesses while critically important for others. In any case, it’s a great reminder to review customer value drivers and ensure pricing is aligned with those drivers (especially when B2B spending is generally contracting).
For the hybrid model, how do companies strike the balance between subscription and usage? When it comes to pricing GenAI features for SaaS products, we believe companies must consider the role of GenAI within the offering. If GenAI is integral to the product, usage-based pricing becomes a natural fit. On the other side of the spectrum, if the GenAI feature is more of a complement to the core product and the risk is low as it relates to scaling of COGS, companies could consider a standard, non-usage-based pricing scheme.
Where GenAI variable cost is core to the product offering, companies can explore “per query” or “per result” pricing. The “per query” approach reflects the OpenAI model, charging customers based on their usage of generative AI capabilities. The granularity of counting, whether per token, per letter, per word, per query, or per conversation, offers transparency and flexibility, directly correlating usage with value extracted from the solution.
The “per result” approach links price to specific tangible business cases. Examples may be pricing per resolved case (for support software), per campaign launched (for marketing solutions), or per person hired (for HR systems) – tethering the cost to tangible business value. Results-based pricing that is closely aligned with customer value drivers can sometimes be more predictable. For example, a customer support team likely knows their typical case volume and can make estimates as to how many “results” will be addressed by an AI solution.
For instances where GenAI variable cost is in play but is not directly correlated with customer value, “capped packages” or “embedded AI premium” are two suggested approaches. In the “capped packages” model, GenAI usage is limited based on pre-defined buckets or packages, charging a recurring subscription for the allotted usage. Once a customer reaches the limit, the GenAI features are disabled or charged in an overage model, providing predictability for both the SaaS provider and the customer, and creating opportunities for upsell.
The “embedded AI premium” approach blends the value of the GenAI capabilities seamlessly with the overall offering. In this case, the vendor will increase the price point, reflecting the enhanced value brought by GenAI – and for the vendor, the associated cost. GenAI is not a separate entity on the price list but an integral part of the improved offering — working behind the scenes to elevate the overall product without altering the pre-existing pricing structure.
Unless inherently built on a GenAI model, most existing SaaS businesses will likely explore hybrid pricing with elements of the “capped packages” model we have suggested. At least for the near-term, I believe this approach will unlock the customer value from GenAI while maintaining P&L balance for the vendor. A few other potential benefits:
- Maintaining subscription-based pricing: By retaining this pricing model at the core of the offering, SaaS companies can continue to provide their customers with consistent pricing and product development/evolution.
- Cost control: Capping usage in a variable cost model helps SaaS providers control COGS. By setting usage limits within the subscription, customers are less likely to overuse the GenAI features, ensuring predictability in costs and preventing unexpected spikes.
- Lever for upsell: By providing additional GenAI usage at a premium cost, SaaS companies can create opportunities to upsell customers who require additional consumption. Assuming that additional usage is not one-time in nature and provides incremental value, it may become the new recurring revenue baseline for that customer.
Pricing GenAI features for SaaS B2B companies is a relatively new consideration, but it can be possible to meet customer demand, deliver incremental value through GenAI, and maintain a balanced cost basis for the vendor.
________________________________________________________
The commentary above has been prepared solely for informational purposes. It is being provided merely to provide a framework to assist in the implementation of the reader of the above commentary’s own analysis and views on the topic or topics discussed therein. Certain information contained therein has been obtained from sources prepared by other parties, which, in certain cases, have not been updated through the date of publication.
The commentary above is not, and may not be relied on in any manner as, legal, tax or investment advice or as an offer to sell any securities or a solicitation of an offer to purchase any securities. The commentary above discusses broad market, industry or sector trends, or other general economic, market or political conditions, and should not be construed as research, investment advice or any other investment recommendation. Market and economic conditions in the future may vary. There can be no assurances that any of the trends or conditions described therein will continue or will not reverse. Past events and trends do not imply, predict or guarantee, and are not necessarily indicative of, future events or results.
There is no guarantee that any strategy or initiatives with respect to computer programs and systems described in the commentary above, including, but not limited to, programs and systems that utilize artificial intelligence, machine learning, probabilistic modeling and other decision science technologies (collectively, “AI”), will be successful. PSG makes no claim that AI is utilized in making investment decisions.
The commentary above may contain projections or other forward-looking statements regarding future events, targets, forecasts or expectations, and is only current as of the date indicated. PSG assumes no duty, or undertakes any responsibility, to update such forward-looking statements.
By accessing the commentary above, the reader acknowledges its understanding and acceptance of the foregoing statement.