Enterprises aren't losing ground on AI because they lack ambition. They're losing ground because many are still clearing debris left by years of incomplete automation: data they can’t fully see, point solutions that don't talk to each other, workflows stitched together with manual handoffs. You can't build autonomous systems on top of that. First, you have to clear the ground.
That's the real story behind a number that should stop executives cold: Despite a 110% surge in AI spending from 2025 to 2026, the average enterprise's AI maturity score is just 51 out of 100, according to the ServiceNow Enterprise AI Maturity Index 2026. More than half of executives (59%) are already using agentic AI. But only 9% have managed to get it to run autonomous, multistep work. The rest are dealing with what the research calls chaos: ambition that has outpaced execution.
But some organizations are clearing the path to AI, and they're saying so in independent public reviews. G2 collects direct feedback from verified software customers, and this year, our customers used those reviews to describe exactly this process, whether it was fixing bad data, retiring disconnected tools, or eliminating governance blind spots.
These are the specific, unglamorous obstacles that stand between "using AI" and "trusting AI to act on its own." Their words are the evidence. G2's recognition of ServiceNow—more than 370 leader badges across 60 categories this year—is a measure of how often, and across how many products, customers point to the same underlying value: ServiceNow clearing the operational obstacles that stand in the way of AI maturity.
Small steps is the analogy that Nisarg Shah, vice president of IT operations management at ServiceNow, uses to frame the problem regarding AI facing many IT departments. Before Neil Armstrong could take the giant leap for mankind during the Apollo 11 moon landing, he first had to trust every small step down from the ladder. The same is true here.
Whether the obstacles live in IT, HR, customer relationship management (CRM), or the automation layer underneath all of them, customers leaving independent reviews in G2 describe the specific things that they had to fix and then trust before they could trust AI with critical business processes.
IT teams don't lack data—they're drowning in it. Alerts, tickets, and infrastructure signals pile up faster than any team can make sense of them, and without a way to turn that noise into a clear next action, even the best AI agent has nothing reliable to act on.
Before ServiceNow, that's exactly the state customers describe. "Our information was scattered across different places, and it was difficult to get a clear picture when something went wrong," says Saloni P., an IT operations user at a large financial services firm.
The fix isn't glamorous; it's just having one place to look. IT Service Management Administrator Kevin R. put it bluntly: "I was never a fan of Excel spreadsheets, and with ServiceNow, I almost never run those types of reports anymore."
For NaveenKumar S., a ServiceNow consultant, the before-and-after was less a report and more a rescue: "ServiceNow ITSM has helped us eliminate the chaos of managing IT requests through emails and spreadsheets. It brings all incidents, service requests, and change management into a single streamlined platform."
And once that single view exists, it stops being just an IT convenience. Product manager Raviya K. describes using the same operational data to make product decisions: "Recurring problems, service performance issues, and visibility into how different parts of the system are connected can highlight where the product or the way things work needs improvement."
That's the definition of actionable intelligence: not more data, but data someone can finally act on. It's also the first thing an AI agent needs before it can be trusted to act on its own.
In ServiceNow’s Enterprise AI Maturity Index research, inadequate data accuracy, access, and management was the most-cited barrier to AI adoption, named by 71% of executives.
G2 users backed that up in their own words, ranking ServiceNow Autonomous IT No. 1 in IT service management, IT operations management, IT asset management, and strategic portfolio management.
On the customer service side, the same disconnect shows up as duplicated effort instead of a customer waiting on a technician.
Sometimes the obstacle is as ordinary as an inbox. Melinda F., an administrator at an IT services firm, describes a support process that lived in one: "Before adopting ServiceNow, our support process depended almost entirely on email. With ServiceNow, customer communication is now tracked in a single platform, which makes it much easier to follow conversations and stay organized."
This pattern also shows up in customer experience: When service history and field context live in separate places, teams lose the thread.
Morimichi S., a telecommunications customer, credits the platform with fixing exactly that: "It helped us integrate a previously siloed customer experience and reform our [employee experience]."
Parichay B., who works with field service at a company of more than 10,000 employees, describes exactly what that guessing looked like before: "Customers would call asking where their technician was, and we'd have to dig through email chains and spreadsheets."
Also on the field side, Ali H. describes the same shift in plain terms: "It is helping the automation of large project tasks which we had struggled managing manually prior to ServiceNow."
Four very different customers (a large field service operation, a telecom operator, an email-bound support team, and a logistics company) hit the same wall: context that lived in someone's inbox instead of the system of record.
Only 16% of organizations have replaced fragmented legacy systems with an integrated platform, according to the Enterprise AI Maturity Index—which is exactly the wall these four customers hit before fixing it.
G2 users backed this up directly: They ranked ServiceNow Autonomous CRM a leader in customer service management, field service management, and configure, price, quote.
Whether you're tracking a laptop or a million-dollar initiative, the underlying problem is the same: You can't hand a decision to an AI agent if the data behind it is scattered or stale, or nobody's quite sure the data is accurate.
That's exactly the state Basavaraju A., an IT asset management administrator frequently experienced before ServiceNow: "Before using it, our asset data was scattered across spreadsheets and different tools, which made it hard to know exactly what we owned, where it was, and who was using it."
Hardit S., a consultant, puts the same problem in terms anyone chasing down a missing laptop will recognize: "Our hardware and software inventories lived in different places, and we often relied on outdated spreadsheets ... I especially appreciate how easy it is to track the full lifecycle of an asset from procurement all the way to retirement without having to chase people for information."
One level up, at the portfolio level, the obstacle isn't missing hardware—it's missing accountability. Jayasakthivel R. says: "The spreadsheet chaos, resource blindness, and the disconnect between corporate goals and technical execution" is exactly what centralizing ideas, funding, and workflows into one data model eliminates.
Yousuf F., describing the same shift in more strategic terms, puts it this way: "Instead of juggling spreadsheets and slide decks, it creates a structured, transparent way to plan, track, and adjust investments so strategy stops being a slogan and starts shaping real execution."
Different altitude, same fix: Replace scattered records with one reliable source of truth. That's the precondition for letting an AI agent make a call on either a laptop or a million-dollar initiative.
Organizations that implement AI initiatives across a wide range of departmental workflows are rare: 61% of Pacesetters have done it, versus just 5% of everyone else, according to the Enterprise AI Maturity Index—the same cross-department scaling problem ServiceNow Strategic Portfolio Management exists to help solve.
HR has a version of the same problem other functions have faced: The answer to an employee's question usually exists somewhere, just not anywhere the person asking can find it.
Guddeti B. describes what that looked like before centralizing: "It eliminates scattered communication (emails, calls, spreadsheets) by centralizing all employee requests into a single system. This improves visibility and ensures nothing gets missed."
Akansh A.'s experience was smaller in scope but the same shape: "The onboarding process was streamlined, which made it easier as a user and reduced the back-and-forth over email."
Describing the change from the HR team's side rather than the employee's, Mohamed A. calls out the bottleneck by name: "Employees can find answers, submit requests, and track their progress without having to chase anyone, which removes HR as a bottleneck problem."
For Mandy W., at a retail company running HR with a small team, the value wasn't about scale at all: "It's providing consistent HR support for a department that is bare bones and unable to manually manage multiple HR functions."
Four different vantage points (employee, HR team, small department, and single process) converge on the same fix: one case record everyone can see. That's the context an AI agent needs before HR can hand it work to do. That recognition holds on the HR side too: ServiceNow Autonomous Employee Experience holds the No. 1 spot in G2's HR service delivery category.
A lot of ink has been spilled lately on whether AI agents can be trusted to stay in bounds and on task. Less has been spilled on a more basic question: Are we giving them problems they can solve?
An agent handed a fragmented case history, an unreliable asset record, or a portfolio tracked across six spreadsheets isn't being tested for judgment; it's being set up to fail before it starts. The trust question everyone's debating is really a readiness question in disguise. And readiness isn't a capability you buy. It's the sum of a hundred unglamorous fixes: the ticket that used to sit in someone's inbox, the license nobody could account for, the case that used to bounce between three systems before anyone could act on it.
That's what makes these reviews worth reading past the star rating. Nobody quoted in this piece set out to write about AI readiness. They were solving a narrower problem—a backlog, a blind spot, a spreadsheet nobody trusted anymore. But taken together, their words describe exactly the groundwork that has to exist before an agent can be handed work to do.
AI is moving deeper into businesses, whether or not that groundwork is in place. Organizations can still decide whether the problems they hand it are ones the technology can reasonably be expected to solve. Judging by their own words, ServiceNow customers believe theirs are.
Find out how ServiceNow can help you connect people, data, and workflows on one AI platform.
