In April this year, ServiceNow announced sweeping changes to how AI products are packaged and offered across the ServiceNow platform. In short, AI products like Now Assist are being embedded throughout the ServiceNow AI platform by default, rather than being purchased through additional licenses. At the same time, AI will be priced by consumption, with AI tools being billed according to tokens or ‘assists’.
Now, partners and practitioners across the ecosystem have a big job on their hands: To understand what impact these changes will have on AI projects and the company’s bottom line. But for that conversation to start, you need to first understand some fundamentals about what the new pricing model will involve and how it will work.
To help with that, we’ve spoken to ServiceNow analysts and AI experts to get under the hood of these changes and work out what’s actually going on here.
How Is ServiceNow Licensing Changing – And Why?
In practice, there are two different changes at play here, which are distinct but related:
- First, AI products like Now Assist will be embedded throughout the ServiceNow ecosystem. A key aspect of this is ServiceNow’s new Context Engine, which works alongside the existing Workflow Data Fabric to create a single unified intelligence layer for AI products.
- ServiceNow products will also now be offered across three main pricing tiers: Foundation, Advanced, and Prime. More fundamentally, there will also be a shift away from the traditional seat-based licensing model, where companies were billed per user/month. Instead, all products will involve a hybrid pricing model, with AI products being charged based on consumption of tokens or ‘assists’.
On the surface, this is designed to make it easier for customers to adopt AI products. The logic goes like this: if AI products are a separate cost, then customers need to make a business case and go through procurement processes before getting started. This essentially creates friction around the adoption of new AI tools, particularly for larger organizations, and discourages experimentation. As IDC analyst Stephen Elliot told Informa TechTarget, the new pricing approach creates a ‘try before you buy’ path for AI tools.
To understand this in more detail, NowBen spoke to ServiceNow AI Specialist Oliver Nowak, who said: “The old licensing model implicitly suggested you needed to be a mature ServiceNow user before you could get value from the AI capability. The new model unbundles that.
“AI is included as standard, which reframes it as an enabler rather than a maturity tier. That’s a positive shift, even if it doesn’t change the day-to-day for most practitioners.”
What Does the Consumption-Based Pricing Shift Mean?
There’s another force at play here as well: The move to consumption-based pricing.
Here’s how it will work: Each licensing tier will include a pool of AI tokens that can be assigned to different AI tools and workloads. Effectively, this lets the customer decide what their AI priorities are and which tools are the best use of the resources available. In practice, straightforward generative AI tasks (e.g., summarizing an email) will require a small number of ‘assists,’ and more complex agentic tasks will require more.
This isn’t happening in a vacuum. In fact, vendors right across the tech landscape are currently moving in a similar direction. Now, SAP, Microsoft, HubSpot, and many others are on a similar journey.
To some extent, this is a response to a phenomenon we’ve been hearing a lot about recently: The SaaSpocalypse. In short, this phenomenon describes concerns around the traditional, seat-based model of SaaS vendors like ServiceNow. The theory is that the pricing model becomes less sustainable as AI agents do more and more work. This has often been cited as a reason why ServiceNow and its SaaS peers have seen muted stock performance in recent months.
But in reality, the shift has been underway for some time. As Alecia Wall, Senior Enterprise AI Industry Analyst, told us earlier this year, “The fact that 50% of [ServiceNow’s] net new business already comes from non-seat-based pricing models tells you they’re not clinging to the old model – they’re already operating in the new one.”
Nonetheless, it’s clear that consumption-based licensing is fast becoming the default setting for AI pricing across the tech market.
When Will the Changes Affect Your ServiceNow Costs?
In theory, the new licensing changes took effect when they were first announced back in April. And for new customers, the new model is likely to be the default.
But a SaaS vendor can’t change its pricing model overnight for existing customers without renegotiating the contract. In reality, the new licensing model won’t take effect for existing customers until contracts come up for renewal. As Nowak explains, “I haven’t seen a customer on the new commercial model yet. I’m still having first-explanation conversations most of the time.”
In practice, this means the changes are likely to be phased in over time for most existing customers. If your contract is up for renewal soon, there’s a good chance you’ll be on the new model sooner rather than later. If not, you’ve likely got some time to work out what the changes will actually mean in practice.
Will Your ServiceNow Licensing Costs Go Up?
This is the million-dollar question, and the truth is that it’s difficult to quantify at this stage.
Like most enterprise SaaS vendors, ServiceNow does not publish pricing publicly – so we wouldn’t expect to see a straightforward breakdown of costs across the new tiers. ServiceNow has also declined to discuss specific pricing details of the new changes.
However, according to the AI Economy, the new pricing is “deliberately set lower than what customers would have paid purchasing those components individually.” Nonetheless, nobody can confidently say at this stage what the changes mean for a customer’s overall bill.
Unsurprisingly, this is creating some apprehension across the community. As Nowak explains, “There’s a fairly consistent worry that the total cost over time ends up higher than what they’re paying today, just spread across a different invoice line.”
How Will This Affect AI Adoption?
If you’re on a per-user/seat licensing model for AI tools, you have fairly free rein to experiment with the technology once you’ve purchased the relevant license. But under a consumption model, that picture starts to look very different. So is there a danger that the new model becomes a disincentive for AI adoption?
Nowak suggests that it depends on the type of customer. “At the enterprise end, I think the new model probably accelerates adoption. Larger customers have the maturity and governance muscle to use ServiceNow in a disciplined way, and bundling AI into the platform as standard removes a procurement step that genuinely mattered”.
This essentially assumes that a lack of additional licensing requirements would offset any disincentives created by a shift to consumption-based licensing.
But Nowak suggests that for other customer segments, the cost/benefit analysis could start to look very different. “The segment I worry about is mid-market. The new model needs customers to be using ServiceNow in a mature way to get the value, and if a customer isn’t likely to use it that way, the calculation starts to look uncomfortable.”
Crucially, this isn’t a reflection of the actual quality of the AI tools – it’s about how high the maturity bar is to actually extract value from them.
What’s the Best Way to Manage AI Costs?
“None of this changes the underlying argument that the customers who extract value from the AI layer in 2027 are the ones doing the unglamorous data, workflow, governance, architecture, and organizational change management (OCM) work…
“If anything, the new commercial model makes that point louder, because the price tag on the AI capability is now visible. The gap between organizations that are ready to use it well and the ones that aren’t has become a budget gap, not just a maturity gap.”
Oliver Nowak, ServiceNow AI Specialist
Nowak has four main pieces of advice for ServiceNow practitioners trying to understand and manage costs moving forward:
- Don’t skip consumption modelling: When scoping out use cases for generative and agentic AI, it’s important to understand what the consumption profile actually looks like. The last thing you want is to build an AI agent or tool without understanding the costs.
- Think carefully about the right pricing tier: By default, your tier choice should sit as close to your projected consumption as possible. Whatever you choose, you need to be aware that it’s a balancing act between your appetite for cost overruns and paying an up-front premium for headroom.
- Skill-level governance is fundamental: It’s important to treat skill-level governance as part of the up-front cost conversation, alongside the raw tokens. This is what stops you from accidentally enabling capabilities you don’t need. This has to be done at the start of the project to manage costs effectively from the start.
- Keep an eye on the dashboards: In ServiceNow, you can get a decent overview of costs through the AI Control Tower, as well as platform analytics in Now Assist and agentic AI usage dashboards. If you’re going to commit to a new model, you need to get your head around these from the start.
Whatever decisions you make, it’s important to properly understand the cost implications and have the data you need to monitor this over time.
Final Thoughts
Ultimately, the licensing changes will bring both challenges and opportunities for those using AI products in ServiceNow. On one hand, it’ll be easier to actually get started with AI tools, since you won’t need an additional license just to experiment. But that change comes hand-in-hand with the need to understand how AI consumption works on a fundamental level.
As Nowak says, “Customers should be using AI where it makes a tangible, material difference, not just for the sake of using it.” As consumption-based pricing becomes increasingly common, the winners will be those who understand where that line is and have the data to back it up.