Facilitator guide

Before participants arrive

1. Keep all five site HTML files together. Open index.html on your laptop.
2. For smartphone access, upload the complete site folder to an approved static host. Confirm that its link opens without a presenter login.
3. Test both activity links on a phone using Wi-Fi and cellular data. Confirm scrolling, county selection, year, growth, offsets, and Back navigation.
4. Create a QR code only after the final HTTPS address is confirmed. Show that address in text alongside the QR code.
5. Keep an extracted offline copy on the presentation laptop and open both dashboard tabs before the session.

Activity 1: What changed? — 5 minutes

1. Minute 1: Open observed revenue changes. Explain state fiscal years and the recipient selector.
2. Minutes 2–3: Participants choose any county, inspect dollars and percentages, and compare recipients.
3. Minute 4: Pair discussion: “What else would we need to understand this change?”
4. Minute 5: Invite two brief observations. Distinguish the measured change from an explanation of its cause.

Suggested debrief: “A reliable number still needs a clear definition, denominator, time period, and decision context.”

Activity 2: What could change? — 7 minutes

1. Minute 1: Open the property-tax dashboard. Identify the historical inputs and conditional assumptions.
2. Minutes 2–3: Choose a county and change lower/central/higher growth. Ask why the dollar loss can remain fixed while its percentage changes.
3. Minutes 4–5: Compare no offset with an illustrative offset. Ask who would authorize and fund that response.
4. Minute 6: Switch to the no-amendment comparator. Identify what it excludes and what it does not model.
5. Minute 7: Ask: “What would you need to verify before explaining this to a resident?”

Suggested debrief: “The value of a scenario is making assumptions visible so people can challenge and improve them.”

Interpretation notes

• Observed half-cent sales-tax distributions and projected property-tax effects are separate analyses.
• State fiscal years in Activity 1 and county fiscal years in Activity 2 are not interchangeable.
• FY2026–27 zero amendment impact is an inherited transition assumption.
• Historical FAC dollar losses are fixed across growth settings. Growth bands are sensitivities, not confidence intervals.
• Local and state offsets are assumptions, not legal capacity or guaranteed funding.
• Duval uses a consolidated-government proxy; comparisons need caution.
• The dashboards are not household tax calculators or predictions of service cuts.
• Do not describe the demonstrations as active Polk County collaboration or county endorsement.

Privacy and fallback

No polling or collection is included. Do not promise anonymous web access without checking the host’s logging practices. If later adding a response platform, check identity/IP collection and avoid names, county identifiers, or identifying free text when anonymity is essential.

If connectivity fails, project the local dashboards and ask pairs to choose assumptions verbally. No participant login is needed.

Activity 3: Fiscal resilience — 12–15 minutes

Open the fiscal-resilience exercise

1. Minutes 0–2: Reconnect to the property-tax scenario. Introduce a hypothetical rural county considering a data center and AI-assisted work. Neither is a guaranteed budget solution.
2. Minutes 2–5: Participants select their first three evidence requests on their phones. They do not submit them. A printed or projected evidence menu works as backup.
3. Minutes 5–8: Pairs explain their choices and identify an omitted question. Invite differing priorities without treating selection counts as a poll.
4. Minutes 8–11: Discuss the resident-complaint example. Connect privacy and security to full costs, review workload, service quality, and trust.
5. Minutes 11–13: Ask what fiscal benefits could be verified and when. Distinguish investment from net county benefit and time savings from cash savings.
6. Minutes 13–15, if available: Each pair proposes one pilot task, one benefit measure, one safeguard, and one county owner.

Fit this within the planned governance/case-debrief allocation; it is not an extra 15 minutes beyond the 90-minute session. The three dashboard exercises together occupy 24–27 minutes.

No named county proposal or current legal claim is made. Do not imply that hosting a data center requires county adoption of AI, or that either follows necessarily from the amendment. Source links provide background only. No opinions are transmitted or aggregated.