A recent piece in the Wall Street Journal dives deep into how AI software vendors have yet to find the perfect pricing model, as businesses hesitate to adopt new tools that don’t always have a crystal-clear return on investment.
Isabelle Bosquet unpacks how, initially, companies priced AI tools with a per-user, per-month fee, much like other SaaS products, but the high cost of computing made these prices steep.
According to Bousquette, “AI’s high compute demands meant they needed to charge eyebrow-raising prices to cover the cost of delivering the service.” This led to pushback from CIOs who weren’t willing to pay a significant premium for AI-powered versions of existing tools.
As CTO of Stagwell Marketing Cloud (SMC), I've faced similar roadblocks. In traditional service businesses, pricing is based on hours worked or fixed contracts, while SaaS products have typically relied on subscription-based or per-seat pricing. AI doesn’t fit neatly into either model.
Calculating the ROI on AI tools can feel daunting, since traditional models of doing so can fall flat. “ROI analysis [for AI applications] involves synthesizing a small mountain of factors,” Manal Yousef notes, “from the complexity of integrating AI systems, to data accessibility and quality, to attribution challenges, to time saved by automating rote tasks.”
In response, vendors are experimenting with new pricing strategies, from bundling AI features into existing software to introducing more flexible, consumption-based pricing or bespoke packages.
At SMC, that might mean experimenting with hybrid pricing models. If a large-headcount company adopts an AI-driven platform, they might be charged a per-seat cost, without total AI usage factored in.
But if a smaller company is the client, it might make more sense to charge them a base subscription fee, as well as usage fees above a certain limit.
This structure ensures predictability for larger enterprises while keeping pricing flexible and scalable for smaller teams or businesses with fluctuating AI needs.
As Bousquette explains, businesses are still trying to figure out the concrete value of AI, which makes committing to traditional pricing models difficult. “Now vendors are making changes to how they price in an attempt to gain more users and more adoption,” she writes, pointing to Microsoft’s shift toward usage-based billing with Copilot Chat.
Harris QuestDIY—a self-serve, AI-powered survey platform that’s part of Stagwell Marketing Cloud—allows users to create and deploy surveys on an individual basis, without necessarily committing to a full subscription.
This can lower the bar for entry for hesitant companies who might want to see actual results before committing to a longer-term SaaS platform, as can various “freemium” offers that give clients a taste of what AI-driven solutions can do for their businesses .
Regardless, pricing remains a major sticking point, and companies are weighing whether it’s worth buying AI tools or building their own. (At Stagwell Marketing Cloud, we have some thoughts on this.)
“Software companies are also facing pressure to adapt their pricing to account for the fact that the actual cost of using the underlying models is going down,” Bosquette writes, a trend that has only increased since the launch of DeepSeek.
From my point of view, educating clients on AI pricing is a process, and it’s not purely about automation and dollars saved. AI should reduce costs, enhance efficiency, replace redundant expenses, and open up entirely new revenue opportunities—not just be another line item in a company’s budget.
What you’re ultimately paying for isn’t just a cheaper way of doing the same rote work. It’s a paradigm shift, and those take time, patience, and buy-in.
AI opens up the door for brand new things—new experiences, new ways to create. That’s where people will pay more money.
