Our CEO told CRN the constraint has moved. Here's what this means for anyone planning an implementation.
Reporting on ServiceNow's second-quarter earnings call, CRN covered the company's plan to sell well below its usual enterprise buyer. ServiceNow Chairman and CEO Bill McDermott told analysts the first of several AI-native products is in beta and near release: "...a conversational service desk experience: no tickets, and [including] AI-coded automation," sold self-serve to a market he termed the "Fortune 500,000."
CRN spoke with executives from several leading partners (including several prominent thoughts from our own Rick Wright), who were largely encouraged by the announcement, agreeing that AI is what was missing when ServiceNow last pushed into the midmarket, while posing a pointed question about how professional services fit a self-serve product.
For 20 years, the critical path of a ServiceNow implementation ran through build effort, involving process design, configuration, testing, the long tail of upkeep, etc. If AI collapses that path, the offering could reveal what was always sitting behind the build: decisions, data, and human behavior.
Rick was one of several partner executives CRN asked about the move:
"It's quickly getting to the point where the limiting factor of how fast we can go in these implementations is how much change the customer can absorb in how short a time frame. It's no longer about how much time it takes to configure the platform. It's really the adoption and the change management that's the limiting factor."
ServiceNow has gone downmarket before, and Rick was direct about why it didn't land the first time:
"What they tried to do [with ServiceNow Express] was just simplify the product without creating a whole other version. At its heart, it was the same platform. It didn't really change the complexity or what you needed to do to support it."
Simplifying a packaging tier doesn't help a 400-person company with no process library, no dedicated platform team, and no appetite for a six-month design phase. The complexity didn't live in the SKU, but rather in what you had to do to stand the thing up and keep it running.
AI-native changes that, and Rick's framing of the mechanism is worth quoting in full because it names each place the effort actually sits:
"Leveraging AI now, you know how you configure the platform, how you support the platform, how you have agentic agents to drive a lot of the workflows and a lot of the configuration, the upkeep, how you do upgrades. AI is completely changing how much effort it takes to support the platform, and I think now you can go downmarket where you couldn't before."
He isn't alone in that read. Ahead's Ryan Crosby told CRN something similar, albeit from a different angle, that ServiceNow has tried this before and that smaller companies, lacking a deep process inventory, will need something closer to an assembly-line approach to deployment.
When two partners with very different practices independently identify the same missing piece, that's often worth more than either observation alone.
There are some legitimate concerns to consider about this move. Konversational's Andrew Paolino indicated that partner businesses are built on delivering professional services, and if there are no services to deliver, the offering isn't interesting, unless it proves to be a door-opener whose customers become strong candidates for the broader platform.
Paolino also cited the competitive pressure of AI-native ITSM entrants promising outcomes without the workflow build and the consulting spend.
Rick was explicit with CRN that this isn't a midmarket-only move, saying, "CoreX has invested heavily in R&D on using AI to drive more efficient and expedient implementations, is already well down that path, and sees it as a need not just for the midmarket but for every client, because they all expect it."
The part nobody outside ServiceNow can answer yet is how this reaches customers. Rick put the uncertainty and the reassurance in the same breath:
"I think what will be interesting is the go-to-market sales channel marketing. We do a lot with ServiceNow today in joint marketing activities and customer roundtables, but we haven't seen their go-to-market model for the upcoming technologies. … but I know ServiceNow is still primarily a channel company."
A product-led motion and a channel-led motion are different machines. How they interlock, determining who sources, who implements, who supports, how a self-serve customer becomes a platform customer, etc., become the variables that will determine how the channel accepts this.
ServiceNow declined to discuss the downmarket move in more detail but confirmed the "Fortune 500,000" descriptor and said more information is forthcoming.
If the platform is going to arrive able to answer questions, configure itself, and act on your behalf, then what a midmarket buyer needs to provide is clarity about their own operation. Four questions determine whether an AI-native service desk will work in your environment:
What data do you need before implementing an AI-native service desk?
Is our service desk data clean enough for AI agents?
What should AI agents be allowed to do without human approval?
Does a self-serve ServiceNow deployment scale into the full platform?
What data do you need before implementing an AI-native service desk?
Before an AI-native service desk can deflect work, you need a documented inventory of your top request types, covering the volume of each, the path a request takes today, and the person or team that owns it. That inventory is the primary input any AI-native service desk consumes, and no amount of platform intelligence substitutes for it.
In our experience the value of the exercise is less in the document it produces than in what it exposes, because the request types nobody can describe are usually the ones absorbing the most time.
Is our service desk data clean enough for AI agents?
An AI agent answers against your systems of record, so the readiness test is whether those records are trustworthy enough to answer a question against. Work through each data domain and establish which system is authoritative, how current its records are, and where duplicate or conflicting entries live.
Data that a human can quietly work around will stop an agent cold, which is why this assessment belongs before the buying decision rather than after it.
What should AI agents be allowed to do without human approval?
Define the autonomy boundary before the agents arrive rather than after. For every task you plan to hand to an AI agent, decide what it may do unattended, what requires human approval, who reviews the exceptions, and how you would evidence any of that to an auditor. Organizations that defer this work usually end up writing their governance during an incident, when the pressure to decide quickly is highest and the record is thinnest.
Does a self-serve ServiceNow deployment scale into the full platform?
Assume you will grow into more of the ServiceNow platform than the service desk you start with. Favor decisions on data structure, naming, and integration that keep a later expansion straightforward, rather than ones that optimize narrowly for the initial footprint. A self-serve entry point delivers its value only when the work you do inside it carries forward.
A buyer who has inventoried their request types, tested their data for authority and currency, and set an autonomy boundary can evaluate an AI-native service desk on its merits, because they will know within a week whether it is answering questions against their actual data.
Bill McDermott's vision is ambitious: a product-led motion to organizations historically outside ServiceNow's target audience. As Rick suggests, its success will depend on whether AI has truly shifted the implementation bottleneck from software configuration to organizational readiness.
But if configuration stops being the constraint, then the challenge in front of every organization considering these ServiceNow offerings is the same as always: knowing what you want, having data worth acting on, and getting your teams adapted to a new way of working.
--
CoreX is a ServiceNow Elite Partner and implementation consultancy. If you're weighing what an AI-native service desk would mean for your organization, or how to get your data and your change plan ready for it, we're happy to have the conversation.