Tessa AI Policy Template for Workplace Operations

A practical Tessa AI policy should define what the assistant may do, what requires human review, which information it may use, and how employees can challenge or correct an output. The goal is not to approve every use of AI in the abstract. It is to create clear operating boundaries for workplace teams using AI to find information, summarize requests, identify patterns, and support decisions without transferring accountability to a tool.

This template is designed for workplace, facilities, employee experience, operations, and people leaders. Adapt the bracketed choices to your organization, then publish the final policy alongside your broader privacy, security, records-management, and acceptable-use policies.

Policy purpose and operating principle

Purpose: This policy governs the responsible use of Tessa AI for workplace operations, including workplace information discovery, request triage, operational analysis, employee support, and preparation of recommendations.

Operating principle: Tessa AI may assist people with workplace work, but it does not replace accountable judgment. A designated owner remains responsible for decisions, communications, approvals, and actions taken in the workplace environment.

The policy should be easy to understand in daily use. Employees should know when Tessa is appropriate, what they should avoid entering, and how to request human assistance. Administrators should be able to review whether the system is being used consistently with organizational policies.

Approved workplace uses

Subject to the organization’s configured permissions and data controls, Tessa AI may be used to support tasks such as:

• Finding or explaining workplace information that the user is authorized to access.
• Summarizing workplace requests, recurring issues, or operational themes for the responsible team.
• Helping employees understand workplace processes, available resources, or request categories.
• Preparing draft responses, announcements, or internal guidance for human review.
• Identifying possible patterns in space, booking, request, or workplace operations data.
• Helping workplace teams organize questions before making a decision or escalating an issue.

These uses are assistive. A Tessa response should be treated as a starting point for investigation or communication, not as evidence that a policy exception, access change, personnel action, or facilities intervention has been approved.

Human review and decision ownership

Human review is required before an AI-assisted output is used to make or communicate a consequential decision. Consequential decisions include changes that may affect an employee’s access, safety, privacy, accommodations, work location, assigned space, visitor access, or employment experience.

The responsible team must verify important facts, check the relevant policy, and confirm that the proposed action is within its authority. For example, Tessa may help summarize a pattern of workplace requests, but a workplace leader should determine whether the pattern warrants a policy change. Tessa may help draft a response, but the accountable owner should confirm that the response is accurate, appropriate, and consistent with employee commitments.

Each organization should name decision owners for common outputs:

• Workplace operations: operational requests, space changes, and service coordination.
• Facilities and safety: physical conditions, emergency matters, and building risks.
• People or human resources: employee relations, accommodations, and sensitive personnel matters.
• Information security and privacy: access, data handling, retention, and suspected misuse.

Information handling boundaries

Users may provide only the information needed for the approved workplace task. They should not enter sensitive personal information, confidential investigation details, credentials, security secrets, or unrelated employee records unless the organization has explicitly approved that use and configured appropriate controls.

Do not use Tessa as a substitute for a secure case-management system. Sensitive matters should be routed through the organization’s approved channel, especially when they involve health information, allegations, legal advice, disciplinary action, security incidents, or personal safety.

Users must not attempt to obtain information by asking Tessa to bypass permissions, infer restricted personal details, reveal another person’s private activity, or combine data for an unauthorized purpose. Access to a Tessa-assisted answer does not expand the user’s underlying authorization.

Use cases that require escalation

Escalate to the designated human owner rather than relying on an AI response when:

• The answer could affect safety, accessibility, privacy, security, or employee rights.
• The request concerns a complaint, investigation, accommodation, health matter, or personnel decision.
• The information appears incomplete, contradictory, outdated, or unexpectedly specific.
• The output recommends an action outside the user’s role or approval authority.
• An employee disputes the information or asks for a correction.
• The request involves an emergency or an immediate threat, where established emergency procedures take priority.

Escalation should be a normal service path, not a failure. The policy should identify where users can go, the information they should include, and who owns the response.

Decision boundaries for common workplace tasks

Decision point	What to evaluate
Assist with routine information	Low-risk workplace information, process explanations, and draft summaries. Human review is useful but not always required before sharing.
Human review required	Operational patterns, draft employee communications, and recommendations that may influence a workplace action. An accountable owner verifies facts and approves the result.
Escalate to a specialist	Safety, privacy, accessibility, security, personnel, accommodation, investigation, or emergency matters. Route to the designated human team instead of relying on Tessa.

Use the following matrix as a starting point for configuring guidance, training users, and reviewing proposed use cases.

Review controls for Tessa AI use

For each proposed use case, document the task, intended users, data involved, expected benefit, human owner, approval requirement, and escalation route. Review the use case before launch and whenever the workflow, data source, permissions, or business purpose changes.

A lightweight review record can include:

• Use case: What workplace problem is Tessa helping with?
• Users: Which roles may use the capability?
• Inputs: What information may be supplied or accessed?
• Outputs: What may Tessa produce, and what may it not decide?
• Owner: Who verifies results and approves action?
• Escalation: Where do uncertain or sensitive cases go?
• Review date: When will the organization reassess the use case?

Connect this review to the organization’s wider connected workplace management platform governance. A clear operating context helps teams distinguish an informational answer from an action that changes a booking, request, assignment, or workplace record.

Employee transparency and recourse

Employees should be told when AI assistance is materially involved in a workplace service or communication, where that is required by organizational policy or applicable law. They should also have a practical way to ask questions, correct inaccurate information, or request human review.

Transparency does not require exposing technical details that would confuse users. It does require explaining the role of the tool, the responsible team, the types of information used, and the route for correction or escalation.

When Tessa helps draft an employee-facing answer, the sender remains responsible for its content. Communications should avoid presenting an unverified suggestion as a final policy, entitlement, or confirmed operational fact.

Monitoring, training, and policy maintenance

Administrators should monitor usage at an appropriate level to identify recurring errors, unsupported requests, permission problems, and opportunities to improve workplace guidance. Monitoring should follow applicable privacy, labor, and information-governance requirements. It should not become a hidden system for evaluating employees’ productivity or personal behavior unless separately approved for a lawful, clearly communicated purpose.

Training should cover approved uses, prohibited inputs, verification, escalation, and the difference between an answer and an authorized action. New users should practice with low-risk workplace questions before using AI assistance in more sensitive workflows.

Review this policy at least annually and after significant changes to Tessa, workplace systems, organizational policies, or applicable requirements. Incorporate lessons from employee feedback, corrections, security reviews, and operational incidents. Teams using workplace requests can use those correction and escalation patterns to identify where instructions or routing need improvement.

Adoption checklist for workplace leaders

Before publishing the policy, confirm that:

• The approved uses are specific enough for employees to apply.
• Prohibited or restricted information is described in plain language.
• Human owners are named for consequential decisions.
• Escalation routes are active and easy to find.
• Permission boundaries match the underlying workplace systems.
• Employees have a correction and human-review path.
• Training and administrator review responsibilities are assigned.
• The next policy review date is recorded.

Frequently asked questions about Tessa AI policies

Does every Tessa response require human approval?

No. Low-risk informational assistance may not need individual approval. Human review is required when an output informs a consequential decision, changes a workplace record, communicates a sensitive conclusion, or creates a commitment on behalf of the organization.

Can Tessa make workplace decisions?

Tessa can support analysis and prepare recommendations, but the organization should retain decision authority with an accountable human owner. The policy should state which actions require approval and which roles may approve them.

How should employees challenge an inaccurate answer?

Provide a clear human escalation route, such as a workplace operations queue or designated service owner. The request should capture the disputed information, the relevant context, and the action the employee needs reviewed.

Policy owner approval block

Policy owner: [Name or team]

Approved by: [Role or committee]

Effective date: [Date]

Review date: [Date]

Related policies: [Privacy, security, acceptable use, records management, accessibility, and workplace policies]