The AI readiness gap is the distance between how much AI an organization wants and how much it is organized to use. New research from Google Cloud and Hypothesis Group shows the gap did not close this year: 91% of organizations have an AI-first goal, but only 42% say it has been translated into clear ways of working (Google Cloud and Hypothesis Group, 2026). Dan Turchin, CEO of PeopleReign, was one of three experts consulted for the study.
This post walks through what the 2,638-person survey found, what the few mature organizations do differently, and what it means for CIOs and CHROs who own employee-facing AI.
What is the AI readiness gap?
The AI readiness gap is the difference between an organization’s AI ambition and its operating readiness: the workflows, skills, governance and budget needed to turn AI adoption into results. Adoption is no longer the constraint. Optimism rose in the 2026 study, yet the gap between ambition and readiness stayed where it was a year earlier.
Dan Turchin put the starting point plainly in the report: “AI is no longer optional.” His point is that nearly every enterprise already has at least one use case that shows what the value could be. What is missing is the organizational structure around that use case.
The study surveyed 2,638 executives and knowledge workers at organizations with 300 or more employees in the U.S., U.K., France, Brazil, India and Japan. Fieldwork ran in August 2026 (Google Cloud and Hypothesis Group, 2026).
Why do AI-first goals stall before they change how work gets done?
AI-first goals stall because organizations are asked to show results before they have redesigned the work. Among executive decision-makers in the study, 87% said they are asked to show productivity gains before workflows are redesigned, 78% said expectations are moving faster than they can be implemented, and 76% feel pressure to show measurable impact before the company is ready to deliver it.
The adoption curve shows the result. The share of organizations still in the initial or exploring stages fell to 63%, and scaling grew from 18% to 27%. The share of mature organizations stayed at 10%, unchanged from 2025. Turchin described the cause in a single sentence: organizations mature more slowly than technology, and “the technology is way ahead of organizational readiness.”
For IT and HR leaders, this explains a familiar pattern. A chatbot launches, usage looks healthy, and ticket volume and employee frustration barely move, because the underlying service process never changed.
What do mature organizations do differently?
Mature organizations redesign the work around AI and put AI inside the tools people already use. They are more likely than organizations at the scaling stage to say AI has fundamentally changed their business processes, to require AI to be built into existing apps, and to align AI with the needs of teams and users. They also see stronger results.
| Measure (2026) | Mature | Scaling |
|---|---|---|
| AI helps complete tasks faster | 63% | 48% |
| Employees are enthusiastic about being asked to use AI | 63% | 44% |
| AI increases innovation | 60% | 48% |
| AI has fundamentally changed business processes | 44% | 34% |
| AI is aligned with team and user needs | 54% | 41% |
| It is important to have AI fully integrated within existing tools and apps | 36% | 22% |
| Companies must move off legacy technology to transform | 85% | 73% |
Source: Google Cloud and Hypothesis Group, Beyond AI Optimism: Closing the Ambition Gap (2026). “Mature” combines the advanced and highly transformed stages.
The integration finding matters for employee service. Employees ask for help in Slack, Microsoft Teams and email, and the work itself lives in systems of record such as ServiceNow, Workday and Salesforce. AI that sits in a separate portal adds a step. AI that works where the request is made removes one.
Are employees ready for AI, and what support do they need?
Employees say they are ready, but they need support to keep up. In the study, 91% of employees feel confident they can adapt to AI-driven change, yet 63% feel pressure to keep up with the pace of AI developments. Organizations see the need: 83% say employee training is a priority, and 60% say they need help keeping employees prepared as tools change.
The research also flagged a side effect that HR leaders should plan for. While 84% of respondents say AI helps employees collaborate and communicate more effectively, 66% say AI creates new silos between people who use it well and people who do not. And 80% agree AI makes human judgment more important, not less.
PeopleReign’s own AI Peer Benchmarks report points the same way. It found that employees with five or more hours of training use AI regularly at a rate of 79%, compared with 67% for those with less (PeopleReign AI Peer Benchmarks 2026, citing BCG AI at Work 2025), and that 56% of HR functions do not formally measure AI success.
Why does trust decide how far AI agents can go?
Trust decides how far AI agents can go because people will not hand over more work than they can verify. AI agents are software that can take actions on a person’s behalf, such as resetting access or updating an HR record, rather than only answering questions. In the study, 97% of organizations use or plan to use them, but 65% are cautious about giving them more tasks, and only 35% of employees say governance for what agents can do independently is clear.
The report quotes Turchin on this: “Innovation can never move faster than the speed of trust.” He said lack of trust is one of the things that constrains agentic adoption. We wrote about the same idea in Agentic AI Doesn’t Matter Until It Gives People Time Back.
The payoff from getting this right is measurable. Highly transformed organizations have agents complete an average of 7.19 unique tasks, compared with 3.37 at organizations just starting out. The widest gap is in using company data and knowledge better: 71% versus 29% (Google Cloud and Hypothesis Group, 2026).
Why can’t organizations predict the cost of AI?
Organizations cannot predict AI costs because usage, models, capabilities and pricing are all changing at once. In the study, 72% said AI goals are expanding faster than their AI budget, 68% worry usage costs could rise faster than expected, and 77% are still learning to manage AI costs at scale. 92% said they need help with AI cost and usage planning.
Turchin’s read on this was direct: “Nobody feels like they have the right inputs or equations” to make these decisions with confidence. Mature organizations are better positioned. The report found 50% of them understand AI costs well enough to plan long term, compared with 35% of organizations at the scaling stage.
The practical answer is to attach cost to an outcome you can count. For employee service, that means cost per resolved request and hours returned to employees, which are easier to forecast than token consumption.
What does closing the AI readiness gap look like in IT and HR service?
Closing the gap in IT and HR service means starting with one high-volume workflow, letting AI act inside the systems employees already use, setting clear limits on what it may do, and measuring outcomes instead of logins. Employee service suits this approach because requests are repetitive, the results are countable, and sensitive cases can go to a person.
PeopleReign is an AI-powered employee service automation platform that layers on top of existing ITSM and HCM systems rather than replacing them. It gives employees back four to six hours of productive time per week by automating common IT and HR requests through ticketless service. Here is how that maps to what the research says mature organizations do:
- AI where work happens: The Virtual Agent works in the channels employees already use and integrates with more than 100 enterprise systems, including ServiceNow, Workday, Salesforce, Slack and Microsoft Teams.
- Clear limits: Guardrails include PII redaction, topic blocklists and role-based access controls, and the Virtual Agent escalates to a live agent when it reaches its limits.
- Measured outcomes: Insights provides predictive analytics for continuous improvement and Experience measures employee sentiment, so results are visible to both IT and HR.
- Fast time to value: The main use case runs in 30 days. A global engineering and technology firm ran a 30-day proof of concept against a goal of 70% accuracy on its own FAQ questions, and the Virtual Agent exceeded it.
Customer results show what scale looks like. A Fortune 500 entertainment company reported human routing accuracy of about 40% before PeopleReign. Since August 2023, PeopleReign Classify has auto-routed nearly 700,000 tickets at over 95% accuracy.
See how this applies to IT service and HR service.
What should CIOs and CHROs ask before the next AI investment?
Ask five questions that test readiness instead of adoption. A yes to each means the organization has moved from using AI to being ready for it.
- Have we redesigned the workflow, or only added AI to the existing one?
- Does AI work inside the tools and channels employees already use?
- Can we forecast the cost of this use case at ten times today’s volume?
- Do employees know what AI may do on its own and what needs a person?
- Are we measuring resolved requests and hours returned, not just logins?
Frequently asked questions
What is the AI readiness gap?
The AI readiness gap is the distance between an organization’s AI ambition and its ability to deliver on it. Google Cloud and Hypothesis Group’s 2026 research found that 91% of organizations have an AI-first goal, but only 42% say it has been translated into clear ways of working. The gap sits in workflows, skills, governance and cost planning, not in access to AI tools.
Why do AI-first strategies stall at the scaling stage?
Most stall because organizations add AI to processes designed for a different way of working. In the 2026 Google Cloud and Hypothesis Group study, 87% of executives said they were asked to show productivity gains before workflows were redesigned. Organizations that move past scaling are more likely to change business processes, integrate AI into existing tools and align it with team needs.
How do you build trust in AI agents at work?
Define what an agent may do on its own, what needs human approval and who is accountable for each action. The 2026 research found only 35% of employees say governance for agent tasks is clear. Start with bounded, high-volume work such as IT and HR service requests, keep humans in the loop for sensitive topics and report results openly.
How should leaders plan for the cost of AI?
Plan for usage-based costs that change as models, pricing and adoption change. In the 2026 study, 77% of organizations said they are still learning to manage AI costs at scale, and 68% worry usage costs will rise faster than expected. Tie spending to measurable outcomes such as cost per resolved request and hours returned to employees, not to logins or tokens consumed.
Where should IT and HR leaders start closing the AI readiness gap?
Start with one high-volume, well-defined workflow where AI can act inside the systems employees already use. IT and HR service requests fit this profile: they are repetitive, measurable and easy to escalate to a person. Set outcome metrics before launch, train employees for their specific roles and expand only when the results hold up.
Read the full research
Want the data behind this post? Go straight to the sources:
