Google announced its Gemini agent for work on October 8, 2026 (Google Cloud). Google describes it as a single agent that can answer questions, handle knowledge work, create content and write code (Google Cloud).
The practical question for a salon, med spa or wellness clinic is not whether to give an agent broad access to your systems. It is whether a limited task could help your team act on first-visit follow-up more consistently. A staff-reviewed list and message drafts are a reasonable place to explore that question.
Why first-visit follow-up is a useful test
A first visit creates a clear point for follow-up: you can check whether the client has made another appointment and decide whether a reminder is appropriate. Staff may already do this manually, but a prepared queue could make it easier to see which records need attention. Treat that as a workflow to test, not a promise that AI will improve retention.
Keep the task specific. Ask the agent to identify first-time clients without a second booking, organize those records into a daily review list and prepare a neutral draft message for staff to edit. For example, the draft could thank the client for visiting and invite them to book again. It should not make claims about what they need or why they have not returned.
Start with an export, not assumed access
Use a clean export that contains only the fields needed for the task: a client identifier, first-visit date, second-booking status and the approved contact details needed for staff to review a draft. Do not assume Gemini connects to your booking software. Google describes Gemini working inside Workspace apps and connecting to software through tools, but whether your particular booking platform is supported and configured needs to be checked (Google Cloud).
Before using any client information, check that the tool and setup are approved for it. For med spas, IV clinics and wellness businesses, leave treatment notes and health details out of this workflow. If HIPAA or another privacy rule applies to your business, use only a setup that meets your obligations. Never send client health or treatment details to ad platforms or general marketing tools.
Give the agent clear boundaries: list the records it found, flag missing or conflicting information and draft messages without sending them. Have staff compare each record with the booking system, confirm the client is eligible for outreach and approve or edit every message. TechTarget’s coverage of the launch notes that agent access raises questions about permissions (TechTarget). The article also quotes an analyst saying that achieving exactness can take time (TechTarget).
Measure the follow-up, not the drafts
Decide in advance what would make the test useful. You might track how many eligible records staff review, how many messages they approve, whether clients rebook within a period you choose and how much staff time the process takes. Also note when staff correct a record or rewrite a draft. A tidy-looking queue is not evidence that the information is accurate or that follow-up is helping.
If you want to learn whether outreach is contributing to changes in rebooking, consider attribution experiments. The point is to evaluate the follow-up work, not just count how many messages an agent can draft. Keep the review process in place while you assess the results.
What to tell your clients
Describe this as a possible administrative aid for preparing first-visit follow-up, not as an agent that independently manages client relationships. Be clear that staff will verify the booking information and approve every message. If your clients ask about data use, explain what information is included, what is excluded and who reviews the output.
What to do
- Choose one location or service line and define which clients belong on the list.
- Prepare a minimal export and confirm the tool is approved for the data.
- Ask for a daily list and draft messages, with no sending permission.
- Review record accuracy, message edits, rebooking and staff time before deciding whether to continue.
Before expanding the test, assess the data, context, access, evaluation and governance the agent would need. Those are the inside-out foundations addressed by Vetraeus Agent Readiness.
- ai-agents
- client-retention
- first-visit-follow-up
- appointment-based-businesses
- data-governance
Alan Jordan, founder and principal solutions architect. Data engineering, data science and business analytics for appointment-based businesses.
