Before you read on, ask yourself:
- Do you know what business information employees are using with Copilot?
- Can your team distinguish between public, internal, and confidential information?
- Does everyone know where to ask for guidance before using AI with sensitive data?
Your business has approved Microsoft Copilot. Employees are excited to use it to draft emails, summarize documents, and get work done faster.
That is a promising start. But before expanding its use, there is an important question to ask:
What information are employees giving it access to?
A request to brainstorm presentation titles is different from a request involving confidential client or financial information. The tool may be the same, but the data involved changes the conversation.
In our recent AI in the Workplace Thought Leadership Session, Kayvan Yazdi and Latifa Afitou explored why understanding your data is an essential part of using AI responsibly. In this article, we build on that conversation and consider the questions businesses should ask as employees begin using tools such as Copilot.
Copilot Is Approved. What Happens Next?
Imagine two employees using an approved AI tool during their workday.
The first asks for ideas for a presentation based on a general topic. The second wants help analyzing a spreadsheet containing confidential client information.
Both employees are using an approved tool, but they are working with very different information.
Before proceeding with the second task, the employee needs to understand the company’s guidance: Is this use permitted? Is the information appropriate for the tool and account being used? Who should review the request if the answer is unclear?
These questions apply to Copilot as well as other AI tools. Choosing a platform is an important first step. Businesses also need to define how employees can use it with different types of information.
Why it matters: Tool approval gives employees a place to start. Data guidance helps them make better decisions about specific tasks.
1. Understand the Type of Information Involved
One practical way to begin is to discuss information in three broad categories.
Public information
Public information is intended to be shared openly. An employee might use a general topic or published material to brainstorm ideas in an approved tool.
Internal information
Internal information is business information that may not be intended for the public. Employees need clear guidance on whether and how it can be used for a particular AI task.
Confidential or regulated information
This may include client records, financial information, employee data, health information, intellectual property, or other sensitive information.
These uses require greater consideration. Employees should understand the organization’s rules and know when to seek guidance before entering, uploading, or making that information available to an AI tool.
Why it matters: The sensitivity of the information is an important part of deciding whether an AI use is appropriate.
2. Ask What Information Copilot Can Access
When evaluating Copilot or another AI feature, it is natural to start with what it can do. Leaders should also ask what information is involved in the intended use.
That includes information an employee enters or uploads, as well as information the feature may access through the applications and accounts being used.
The answer depends on the specific tool, its configuration, the employee’s permissions, and the task. Your IT and security team can help you understand the setup and determine what should be reviewed.
Start with questions such as:
- What information does the employee need for this task?
- Where is that information stored?
- Who currently has access to it?
- Is the intended use consistent with the organization’s guidance?
Why it matters: Understanding the information available to an AI tool helps leaders make decisions about where and how it should be used.
Need Managed IT Services?
We are an Award-winning IT Provider and Comprehensive IT Solutions in San Francisco, San Jose, and throughout the Bay Area.
Schedule A Free Consultation
Need Managed IT Services?
We are an Award-winning IT Provider and Comprehensive IT Solutions in San Francisco, San Jose, and throughout the Bay Area.
Schedule A Free Consultation3. Connect the Rules to Everyday Work
Employees may understand that they should protect confidential information but still be unsure how that applies to a particular task.
For example:
- Can I use Copilot to help draft a general internal announcement?
- What should I consider before summarizing a business document?
- Is client information involved in the task I want help with?
- Who can help me determine whether this use is approved?
A useful policy gives employees direction they can apply during their workday. It should explain the approved tools, the types of information they can use, and where to go when they have questions.
Why it matters: Employees are better equipped to follow the rules when they understand how those rules relate to their work.
4. Give Your Team a Way to Ask Questions
AI features and business needs will continue to change. Employees will encounter situations that are not covered by an example in the policy.
Make it easy for them to ask before moving forward, particularly when they are unsure about:
- The sensitivity of a document or spreadsheet
- Whether a particular use is approved
- What information an AI feature can access
- Whether a proposed task needs additional review
A manager, IT or security team, or another designated person can help determine the next step.
Why it matters: A clear place to ask questions helps employees avoid making assumptions about sensitive information.
5. Build Guidance, Training, and Oversight Around the Tool
An approved AI platform is one part of a broader approach. In the webinar, we discussed six guardrails that work together:
- Approved tools
- Clear data rules
- Appropriate access
- An AI policy
- Employee training
- Ongoing oversight
These guardrails support useful AI adoption. They help employees understand what is available, what information needs protection, and where to get help.
Training can start with familiar examples: brainstorming a presentation title, drafting a routine message, or considering whether a document contains information that requires more care.
Why it matters: Employees need more than access to an AI tool. They need guidance that helps them use it with confidence.
Start With One Conversation
If your business already uses Copilot, start by asking employees and department leaders:
“What are you using it for, and what information is involved?”
Listen for tasks where the value is clear, as well as situations where employees are uncertain about the data. Use those questions to identify where your guidance, access review, or training may need attention.
You do not need to address every possible use at once. Start with the work your team is already doing.
How TruAdvantage Helps
Book Your Free Consultation Now
Copilot and other AI tools can help your team work more effectively. Before expanding their use, make sure you understand the information involved and that your employees know the boundaries.
https://www.truadvantage.com/educational-webinars/
Kayvan Yazdi Kayvan Yazdi, Co-founder and CEO of TruAdvantage, has over 25 years of experience in IT and Cybersecurity. With an MBA in Technology Management from Santa Clara University, he helps California and Bay Area's SMBs and nonprofits build secure, compliant, and scalable IT strategies. A speaker, author, and contributor, Kayvan writes for publications such as Modern Biz IT and the Cybersecurity Bulletin and has been featured on multiple podcasts and webinars. He also serves as a Channel Focus Panel Member and National Tech Day representative for the Bay Area. What he loves most about TruAdvantage is its fun, humble culture, a team that’s always learning, and making clients truly happy.
CEO, Co-Founder
Categories: Blog
