Only 4% of customer service organizations have achieved the customer service AI ROI they expected, according to a Gartner survey of 275 service and support leaders conducted in July through August 2026. Gartner released the figure as its Customer Service & Support Conference opened in Denver. Christopher Sladdin, a senior director analyst in the firm’s customer service practice, used the opening keynote to set out how “AI-forward” organizations differ.
Key Takeaways
- Only 4% of service teams achieved the AI ROI they expected.
- Gartner surveyed 275 service and support leaders in summer 2026.
- AI-forward teams spend nearly 1.8 times more on data foundations.
- Service AI spending rose 38% while overall budgets grew 2%.
Customer Service AI ROI Misses the Forecast
The 4% result is stricter than a test of whether any AI pays. Customer Experience Magazine reports that a separate Gartner survey published in July found only 24% of service leaders “demonstrated positive financial returns across their AI use cases.” Some AI covers its cost. Far less matches the business case.
Measurement is a key part of the story. CXM, citing Sprinklr’s Show Your Work research, reported that just 4% of organizations with AI in live operation had defined success metrics, recorded a baseline, and named an outcome owner. Cost is the other constraint on the record: Gartner has forecast that GenAI cost per resolution will exceed $3 by 2030, more than many B2C companies pay offshore human agents.
What Separates AI-Forward Service Teams
In the Denver keynote, Sladdin described three habits among organizations Gartner calls AI-forward. They aim AI at customer value, invest in connected portfolios and data foundations, and redesign teams.
- Customer value, not only cost. These teams treat interactions as strategic assets and mine conversations and feedback for recurring friction.
- Connected portfolios. They tie a smaller set of use cases to business objectives instead of scattered, one-off projects.
- Redesigned teams. They rebuild roles around what humans and AI each do well, rather than bolting tools onto old job descriptions.
AI’s superpower is that it enables leadership to learn from customer interactions at scale.
Sladdin said that in the Denver keynote, Customer Experience Magazine reported. Gartner predicts that by 2028, organizations that build AI portfolios will achieve twice the business value of those pursuing standalone use cases. He urged “fewer, higher-value and interconnected use cases” and said AI-forward organizations invest nearly 1.8 times more in data and knowledge management foundations than in the AI use cases themselves.
IT-Online’s account of the survey carried the same charge, including his line that humans and AI are not fit for the same job, and that each can deliver value in different ways. He added:
Leaders must stop asking how they remove humans from interactions and instead ask where their involvement creates disproportionate value for customers.
Service AI Spending Outruns the Budget
In August, Gartner reported that service AI spending had risen 38% while overall service budgets grew 2%. AI is taking a larger share of a function whose total budget is barely moving. The figures below come from separate Gartner statements and from Sprinklr research cited by CXM, and they should not be added together.
| Signal | Figure | Context in the reporting |
|---|---|---|
| Expected AI ROI achieved | 4% | 275 leaders, July-August 2026 |
| Positive returns across AI use cases | 24% | Separate survey, published in July |
| Service AI spending increase | 38% | Gartner, reported in August |
| Overall service budget growth | 2% | Same August Gartner report |
| Live AI with metrics, baseline, and owner | 4% | Sprinklr research, via CXM |
| Foundation spend versus AI use cases | Nearly 1.8x | AI-forward organizations |
| Portfolio value versus standalone cases | 2x by 2028 | Gartner prediction |
| GenAI cost per resolution | Above $3 by 2030 | Gartner forecast |
What Service Leaders Should Revisit
For a CFO review, the more useful figure may be 1.8 rather than 4. If the organizations Gartner highlights spend nearly twice as much on data and knowledge as on the AI use cases, a business case that parks knowledge cleanup in a footnote is built to disappoint. That work is slow, and it does not look like AI on an invoice.
The tool choice changes with the portfolio advice. A chatbot that cannot share knowledge with analytics and agent-assist tools will struggle to compound its own gains. Simon Hall, a CCW Europe analyst, told CXM in August: “Organisations don’t scale AI use cases; they scale AI operating models.” CXM noted that platform vendors will like an argument that tends to favor suppliers selling several connected tools, while buyers can still ask how open those connections are.
Workforce design is the remaining constraint. Asking where human involvement creates disproportionate value, rather than how to remove people, undercuts cases built only on deflection. The board question for a large service organization is whether the portfolio, the knowledge base, and the team can carry a return already promised. Related coverage looks at AI payoff outside the tech sector and at Forrester’s AI cost-squeeze outlook.
