AI policies don't have to be a list of restrictions. A framework of AI Zones provides destination organizations a practical way to encourage experimentation while applying stronger controls as data sensitivity, business impact and automation increase.
AI is moving quickly inside destination organizations. Employees are experimenting with tools such as ChatGPT, Claude, and Microsoft Copilot. Additionally AI capabilities are increasingly appearing in our enterprise applications such as CRM, ERP, and other productivity platforms.
At Destinations International’s recent Business Operations Summit, we described this moment as driving a very fast car while still building the guardrails. The goal of AI governance should not be to slow the car down. It should allow organizations to move quickly while putting appropriate guardrails around data, risk and accountability. One practical way to do that is through AI Zones.
What are AI Zones?
AI Zones enhance traditional AI policy into a framework employees can more easily apply to their everyday work. Instead of treating every use of AI as carrying the same level of risk, organizations establish zones based on factors such as:
- The sensitivity of the data
- The AI tool being used
- The business impact of the output
- The level of human review required
- Whether AI is assisting a person or taking an action itself
There is no single AI Zones model that will work for every destination organization. The structure should reflect each organization’s data, technology environment, risk tolerance and culture. A practical starting point is a simple three-zone framework that increases governance as the sensitivity of the data, impact of the work and autonomy given to AI increase. Think of it like a traffic signal: green creates room to explore, yellow calls for additional safeguards and review, and red signals that the use requires significant oversight or may not be appropriate at all.
Here is how that model could work in practice:
Green Zone: Explore and Create
Low-risk uses involving non-sensitive information.
This is where organizations should give employees room to experiment. Appropriate uses might include brainstorming, creating outlines, drafting copy, summarizing public information or testing different approaches to a problem. AI tools utilized in this zone could include free public tools or tools being used in a pilot phase.
The important boundary is data. Confidential organizational information, CRM data, personnel information, licensed datasets and other sensitive information should not be entered into public AI tools. The cultural message of this zone should be: experiment, learn and create within these boundaries.
Yellow Zone: Govern and Review
AI use involving organizational data, higher-impact work or content requiring additional review.
Examples could include analyzing CRM information, developing internal reports, working with financial or performance data, or preparing important external-facing content. These activities may be appropriate, but the tools and processes matter.
Define which approved AI environments staff can place internal data analysis and create approval requirements as AI-generated work moves from internal drafts toward more consequential organizational or external content.
Licensed data requires particular attention. DMOs routinely work with data from providers and organization's right to access data does not necessarily include the right to put raw data into an AI system. Contractual restrictions continue to apply.
In this zone, the message becomes: use AI, but only in an approved environment and with appropriate human oversight.
Red Zone: Control and Escalate
High-impact, sensitive or automated AI uses requiring significant oversight.
This can include AI that takes actions rather than simply makes suggestions, works with highly sensitive information or has the ability to affect employees, partners, visitors or organizational systems.
Examples of work in this zone might include operational automation zone or high-impact, public-facing AI-assisted work. Some uses may simply be prohibited, particularly those involving sensitive HR information, inappropriate use of personally identifiable information, raw licensed data or autonomous decisions that should remain with people.
The message here is: stop, assess and obtain the appropriate approval before proceeding.
Making AI Zones part of the culture
The real value of AI Zones is not the labels. It is making responsible AI easier for employees to understand.
DMOs considering this approach can start with five steps:
- Inventory current AI use. Find out which tools employees are already using and what they are using them for.
- Classify your data. Identify public, internal, confidential, personal and licensed information, and assign ownership.
- Define your zones. Clearly explain the tools, data, activities and approval requirements appropriate for each.
- Create a simple reference guide. Employees should be able to determine their zone without rereading the entire AI policy.
- Train and revisit regularly. AI governance must evolve as tools, vendor capabilities and organizational use cases change.
Most importantly, maintain human accountability. As we discussed at the Summit, the goal should not simply be human in the loop, but human in the lead. People define the objective, select the appropriate data, evaluate AI's work and remain accountable for the outcome.
An effective AI policy cannot simply be a list of things employees are prohibited from doing. That can unintentionally push experimentation underground and contribute to unwanted shadow AI.
AI Zones offer a different approach. They answer three simple questions:
- Where can I experiment freely?
- When do I need additional guardrails?
- When do I need to stop and ask for help?
That makes responsible AI easier to understand and easier to practice. The goal is not to put the brakes on AI adoption. It is to build the guardrails that give our people the confidence to accelerate responsibly.
Share Your Expertise
Want Your Organization Featured Next?
The article you just read was written by someone like you. Submit a blog, case study or social media feature and get your organization's story in front of the DI community.
Join 50+ destination leaders who've shared their stories on the DI Blog