On June 18, the world's largest IT consulting firm, Accenture, released its Q3 results for fiscal year 2026, with revenue of $18.72 billion—slightly above the midpoint of guidance, but Q4 outlook fell far short of Wall Street expectations. Its stock price plummeted 17.2% that day, closing at $129.20. Within two days, its market value evaporated by approximately $14 billion.
On the same day, IBM fell by 6.9%, and Infosys and Cognizant followed suit. The entire IT services sector was jolted awake.
I carefully reviewed Accenture CEO Julie Sweet's remarks during the earnings call. She stated that the company is seeing "more large-scale AI transformation projects" and that demand remains strong. However, she simultaneously cut the full-year revenue growth guidance from 3%-5% to 3%-4%. In simple terms: large projects exist, but small and medium-sized projects are shrinking.
-19% Accenture single-day decline -6.9% IBM same-day decline 91% CIOs don't understand AI dependency 71% difficulty switching AI vendors
The foundation of the per-person-per-day billing model is shaking
Accenture's core business model is "per-person-day billing" — charging based on the number of days an engineer works. This model has operated for decades, stable, predictable, and highly profitable. However, the implementation of AI agents is fundamentally changing this from the ground up.
When an AI agent can replace 3-5 junior consultants in completing "person-day" tasks such as data cleaning, report generation, and process mapping, clients naturally ask: why should I still pay per head?
I noticed a number: Accenture's Q3 order value unexpectedly dropped by 2%. This 2% is not a random fluctuation; it reflects customers reassessing the cost-effectiveness of IT outsourcing. Large projects are still being signed, but traditional "grunt work" type small orders are decreasing—precisely the orders that can be replaced by AI agents.
Core Contradiction: AI agents are not replacing "high-end consulting," but rather the standardized execution work that constitutes 40% of the IT service industry's revenue. The foundation of the per-person-day billing model is being hollowed out.
The IBM report said something even more disheartening
Almost at the same time as Accenture's financial report, the IBM Institute for Business Value released a survey report covering executives from multiple industries worldwide. I read the conclusion several times:
| Research findings | Proportion | My interpretation |
|---|---|---|
| Incomplete understanding of AI dependencies at the vendor, model, and infrastructure levels | 91% | The vast majority of enterprises are in a "black box state" regarding the AI supply chain |
| Believing that switching primary AI vendors or models would be very difficult | 71% | Once chosen, it's hard to switch; vendor lock-in has become a reality. |
| Worry about losing control over AI systems | 62% | The sense of control over data, models, and infrastructure is slipping away. |
IBM Executive Vice President Ana Paula Assis said something I think is underrated: "AI brings new forms of dependency, and losing control means profit pressure or business disruption." In other words: You think you're using AI to reduce costs and increase efficiency, but in reality, you're digging yourself an increasingly deeper supplier hole.
This is not an ordinary corporate research report. IBM is telling CIOs one thing: your understanding of AI systems is far lower than your reliance on them. The gap between the two is risk.
Three Crossroads in the IT Service Industry
I put together Accenture's financial report, IBM's research, and recent industry trends, and came to three conclusions:
1. The "brick-moving layer" of traditional IT outsourcing will be eaten by AI agents
Data cleaning, report generation, process documentation, and basic testing—these standardized execution tasks, which account for approximately 40% of IT service revenue, can already be replaced by AI agents at a rate of 70-80%. The remaining 20-30%, which require human judgment and creativity, will be the value anchor of future IT services.
Accenture has also seen this trend. Sweet specifically mentioned AI transformation projects during the earnings call, essentially telling the market: we are also transforming, shifting from selling headcount to selling AI-driven change solutions. But transformation takes time, and investors are impatient.
2. AI vendor lock-in is the biggest hidden risk for CIOs in 2026
71% of enterprises find it difficult to switch AI vendors, and this is not an exaggeration. Think about it: you choose a large model platform, feed it business data, fine-tune an industry-specific model, deploy intelligent agents, and conduct compliance reviews—only to find the costs are too high or the results are unsatisfactory. Want to switch? Data migration, model retraining, agent reconstruction, and compliance re-evaluation—each of these tasks requires months of work.
Recommendations for CIOs: Before selecting an AI vendor, do three things first — ① require the vendor to commit to data exportability; ② design a contingency plan for model switching; ③ include AI governance tools (such as Microsoft Agent 365, Google AI Control Center) in the procurement list, rather than remedying issues afterward.
3. "AI services" will become a new growth driver — but not through per-person-day billing
IBM fell 7%, but IBM repeatedly emphasized in its earnings report that its full-year revenue growth is still expected to be above 5%, with free cash flow increasing by approximately $1 billion year-over-year. Why? Because IBM's software and infrastructure businesses are growing, especially Red Hat hybrid cloud and AI governance products.
A comparison makes it clear:
Traditional IT Service Model
Revenue growth rate: 3%-4%
Billing method: person-day/project
Growth driver: selling more headcount
AI Impact: Replacing Executive-Level Work
AI-driven service model
Revenue Growth Rate: 5%+
Billing method: Subscription/Token/Effect
Growth Drivers: Selling AI Solutions
AI empowerment: improving delivery efficiency
The growth rates of the two routes differ by 2 percentage points, and their billing logic is completely different. Future IT service companies will either become AI solution providers (selling platforms + governance + transformation consulting) or maintain the traditional model but face an increasingly lower revenue ceiling.
Practical recommendations for heads of enterprise digitalization
This week's events are not isolated stock price fluctuations; they reflect the structural adjustment of the IT service industry under the impact of AI. If you are a corporate CIO or digital leader, I suggest you do four things:
First, examine your AI dependency map. List all AI vendors, models, and infrastructure, marking the degree of dependency and difficulty of switching. 91% of executives are unaware of this; if you do it, you will be ahead.
Second, reassess IT outsourcing contracts. If you are still signing outsourcing contracts billed by person-days, require suppliers to provide AI-enabled solutions — for the same work, AI agents plus human review should reduce costs by over 30%. Suppliers that do not provide AI solutions will be eliminated within the next two years.
Third, incorporate AI governance into the budget. Tools like Agent 365 and AI Control Center are not luxuries but necessities. AI deployment without a governance system incurs compliance rectification costs 6.8 times higher than with a governance system (IDC data).
Fourth, pay attention to the transformation signals of IT service providers. Is the implementation partner you choose still selling headcount, or has it shifted to AI-driven delivery? This judgment directly affects the cost and effectiveness of your project.
