Platforms

ServiceNow falls 27%, its worst day ever, as Gartner finds 97% of companies can't turn AI into savings

ServiceNow fell 27% on October 29 after its results revived AI-disruption fears, and Gartner's buyer-side research, reported a day earlier, had already put numbers on the doubt: 97% of organizations cannot turn AI productivity into cost cuts, yet 85% plan to spend more next year. The savings are late because the freed hours keep being spent the old way.

Richard TangRichard Tang · 3 min read
Share
Gina Mastantuono, ServiceNow's president and chief financial officer, standing with arms crossed
Gina Mastantuono, ServiceNow's president and chief financial officer. On the Q3 2026 earnings call she said AI deployment help on large deals books as services revenue until the contracts renew.

ServiceNow−28.02% — ServiceNow, down 28.02 percent today, the enterprise-software company, sank 27% on Thursday, its worst trading day ever, after third-quarter results revived fears that AI will upend the software business 2. A day before the fall, Gartner+2.69% — Gartner, up 2.69 percent as of its last quote, the technology-research firm, had put numbers on the buyers' side of the same doubt: organizations are creating free time with AI and failing to spend it differently 13. The bottleneck has moved from what the model can do to how work is organized around it.

Boards expect savings, and the budgets rise anyway

Three in four board members expect a productivity-and-savings return on their company's AI spend, Gartner finds, and 97% of organizations cannot cut expenses through productivity measures alone 13. "Ultimately, organizations will not generate cost savings from AI at this point," said CV Viverito, a senior director analyst in Gartner's HR practice 1. Thirty-eight percent of companies are already over budget on AI 1. Eighty-five percent plan to spend even more next year 1. About 95% of the AI-productivity talk on corporate earnings calls concerns future gains rather than results 1.

97% cannot turn AI productivity into cost cuts, yet 85% will spend more next year

0%20%40%60%80%100%Cannot cut expenses via productivity alone97%Plan to spend more on AI next year85%Boards expect a savings return75%Already over budget on AI38%
Data
Value
Cannot cut expenses via productivity alone97%
Plan to spend more on AI next year85%
Boards expect a savings return75%
Already over budget on AI38%
Share of organizations and board members in Gartner's buyer-side research, reported October 28 and 29, 2026. Sources: HR Dive via Yahoo Finance; TechRadar via Yahoo Tech.1,3

The freed time goes back where it came from

The failure has a mechanism. In a Gartner survey of more than 1,800 employees, just 21% of the time AI saved was redirected into more innovative work 1. Creating capacity is not the struggle, Viverito said; what organizations struggle with is making that capacity usable and pointing it at something valuable 1. The freed hours flow back into the same work.

ServiceNow's accounting records the missing work

The quarter was not the problem: revenue matched estimates, net new AI contract value grew 65%, and the company raised its full-year AI target to $1.7 billion 24. What broke the stock was the wait.

Fourth-quarter contracted-revenue growth was guided at 18.5%, below the third quarter's pace and Wall Street's estimate, and Citi+1.52% — Citi, up 1.52 percent today analyst Tyler Radke said investors wanted a sooner payoff from ServiceNow's forward-deployed engineers, who embed inside customers to build AI solutions 245.

Deployment help lands in services, and returns at renewal

The missing work shows up in the accounting: deployment help on large AI deals is booked as services rather than subscription revenue, President and Chief Financial Officer Gina Mastantuono said, and returns to subscription when those contracts renew 6. Salesforce−5.28% — Salesforce, down 5.28 percent today and Workday−4.10% — Workday, down 4.10 percent today fell with it 4; we tracked this doubt into the AI cloud names on October 14, and the buyers' own survey is the new evidence 1.

Three years of roughly 20% growth, then the worst day in ServiceNow's history

$2B$2.5B$3B$3.5B$4B$4.5BQ3 '23Q1 '24Q3 '24Q1 '25Q3 '25Q1 '26Q3 '26Worst day ever followed this print
Data
Total revenue
Q3 '23$2.29B
Q4 '23$2.44B
Q1 '24$2.6B
Q2 '24$2.63B
Q3 '24$2.8B
Q4 '24$2.96B
Q1 '25$3.09B
Q2 '25$3.22B
Q3 '25$3.41B
Q4 '25$3.57B
Q1 '26$3.77B
Q2 '26$3.99B
Q3 '26$4.11B
Total revenue, USD billions, by fiscal quarter. Q3 2026 reported after the close on October 28, 2026; earlier quarters from ServiceNow's SEC filings.2,4,7

Watch the 2027 renewals, and start the redesign now

The same chain has run further inside ServiceNow: its internal use of AI now generates more than $1 billion in annualized value, more than double the level at the end of 2025, Mastantuono said on the call 6. That is what capture looks like after the work is rebuilt around the tool.

The date to watch is 2027, when deployment-heavy contracts renew and the services drag returns to subscription growth 46. Brian Schwartz, an analyst at the brokerage Oppenheimer who recommends buying the dip, conceded the AI momentum "counters a disruption thesis" without establishing a path to reacceleration 5. For the buyers holding the bill, the move is concrete: pick work you understand, put AI on it, and rebuild the work around the hours it frees.

Get Platforms by email, free

The cloud, data and developer software AI is built on. In our daily email, every weekday.

Loading IT

Reader comments

0 comments

    More in Platforms

    All in this topic →

    Platforms

    Cloudflare now bills agent web search, and Exa priced for a buyer who cannot switch

    Cloudflare's product managers set out the terms of the reseller it launched on October 2: one bill at providers' list prices, no markup, and a crawler rule every supplier answers to. Exa's October 18 price sheet reads as if it were written for a buyer who cannot switch.

    Platforms

    FactSet pipes Nasdaq Private Market cap-table data into its terminal before the November 10 Investor Day

    An October 20 deal pipes auditable Nasdaq Private Market cap-table data into the Workstation with no disclosed terms, just as FactSet's subscription growth hits a three-year high. The November 10 Investor Day decides whether it becomes revenue.

    Platforms

    Snowflake priced its coding agent at $6.34 a trial, and the mode that won stays internal

    Snowflake's October 27 benchmark, built from 65 of its own merged pull requests, discloses for the first time what an AI coding agent costs per task inside its product, and shows the winning mode is not for sale. The December print now has to answer whether Snowflake can bill for the outcome rather than the compute behind it.

    Platforms

    Oracle now prices and bills Snowflake on OCI, and owns the discount

    Snowflake gets Oracle's channel and cloud credits; Oracle gets the customer, the contract and the discount on a service that competes with its own database. The case both sides cite is an egress saving for customers they already share.

    Platforms

    Agents launch 17 million databases a day on Databricks; the line they feed makes $100 million

    A new a16z essay adds the first usage numbers under Databricks' $190 billion August mark: agents now out-query humans on the platform and launch 17 million databases a day. The agent-native product line they feed is about 1.4% of a $7 billion run rate.

    Platforms

    Accenture CEO Julie Sweet bought $3.0 million of stock five days after a co-CEO sold $1.4 million

    Julie Sweet put $3.0 million of her own money into Accenture on the open market on October 26, and director Paula Price bought alongside her, days after Asia Pacific co-CEO Ryoji Sekido sold. The purchases land three weeks before the fiscal first-quarter report that tests whether AI price cuts are reaching Accenture's revenue.

    Get the newsletter

    The day's AI and market news, in your inbox every weekday.