EMPIRE OPS × FUTURE CULTURE DESIGN CHECK

De-Identified Quarterly Client Reporting

Key considerations

Get the Zanda data discipline right now to support scale

Everyone uses Zanda consistently: every EAP appointment has the correct Service, Session Pack and final Appointment Status.
No past appointment stays “Pending”: practitioners complete their records and a weekly admin check catches anything unresolved.
Reporting depends on clean source data: unresolved appointments block the quarterly report rather than being guessed or manually reconstructed.

Overview

1. Session correctly recordedZanda statuses mandatory→2. Zanda “Sales” export→3. Export cleanedDeduplicate same Client + Practitioner + Date/Time→4. Export imported to ChatGPTClient + reporting period→5. Report generatedEloise approves

A. Detailed future-state reporting flow

Green · agreed direction / what needs to happenOrange · more work / testing / confirmationGrey · client input / decisionRed · blocked / unresolved

TechDESIGN DIRECTION

1. Set up the EAP customer

The business buying the EAP, e.g. “Entropico”, is set up in Zanda as the relevant Third Party / Sale To, with the organisation-specific Service.

Q01Client question · Calendar-year model

Can we design the new reporting workflow directly for the intended calendar-year model from 1 January, rather than recreate historical client-specific reporting years?

Q02Client question · Live Reporting Register

Are you happy for Future Culture to maintain a simple live Google Sheet Reporting Register with each EAP customer's reporting setup and annual session allowance, updating it when a new customer or contract change is set up?

Contracted allocation

The reporting setup also needs the customer's contracted annual allocation, whether agreed as a fixed number of sessions or calculated from staff/member numbers × sessions. The agreed allowance is maintained in the Reporting Register.

Evidence

Future Culture's organisational EAP customers are present as Third Parties. The Sales report can filter by Sale To, and organisation-specific Services are configured.

HumanPROCESS RULE

2. Book and complete appointments correctly

When booking/delivering an EAP appointment, practitioners apply the correct Service and Session Pack. After the appointment, they apply the correct final Appointment Status.

Q03Client question · Which statuses count as usage?

For allocation reporting, should “sessions used” include appointments marked “Completed” plus chargeable “Late Cancellation”, and any other chargeable status such as “No Show” if applicable?

Q04Client question · Practitioner use of Zanda

Are all reportable practitioners expected to use Zanda consistently? We observed organisational appointments for Eloise and Brett but not Virginia during our inspection. Is Virginia using Zanda for her Future Culture appointments?

Required appointment data

Service, i.e. “Entropico LAP Session” · Session Pack, i.e. “Continued EAP Zoom Online” · final Appointment Status, i.e. “Completed” for an attended session.

Why?

Quarterly automation cannot make incomplete source records trustworthy.

HumanPROPOSED

3. Weekly admin hygiene

Admin checks past appointments weekly. No past appointment should remain “Pending”; it must be moved to the appropriate final status, e.g. “Completed”, “Cancelled” or “Late Cancellation”. Reporting is blocked until unresolved appointments are corrected.

Human boundary

Admin checks and chases. Practitioners confirm unclear outcomes. Confirm whether Virginia uses Zanda consistently.

TechDIRECT EVIDENCE

4. Run Zanda report “Sales”

Run the Zanda report “Sales” for the selected Sale To, date range, practitioners and required Appointment Status filters.

Evidence

Sales exposes Client, Payable By, Practitioner, appointment date/time, Type and Item. It also includes future appointment dates.

Known limitation

Appointment Status is selectable as a report filter but was not observed as an exported column.

AutomationTESTING REQUIRED

5. Clean data and calculate metrics

Use the exported “Sales” spreadsheet to create the validated reporting dataset: remove non-reporting rows, apply the agreed appointment deduplication rule, then calculate unique people, quarter usage, cumulative usage and next-quarter bookings.

Q05Client question · Unique client definition

Should a person with only a chargeable no-show or late cancellation count in “number of clients using the service”, or should this metric count attendees only?

Candidate cleaning rules

In the exported spreadsheet, retain Type = Service and exclude associated Session Pack rows. Confirm whether this leaves one Service row per appointment. Only if duplicate Service rows remain, test the candidate appointment key: Client + Practitioner + exact Appointment Date/Time.

Calculation control

Calculate report figures only from the cleaned, validated dataset using deterministic rules. AI does not decide which appointments count or calculate authoritative usage figures.

AITESTING REQUIRED

6. Generate the report

Import the cleaned export into the ChatGPT reporting template, enter the structured run fields such as Client and Reporting period, and generate the branded draft report using the matching Reporting Register entry.

Q06Client question · Booking reasons

How are the aggregate reasons for session bookings in the current reports produced today, and what source is used?

Stop conditions

Missing configuration, unresolved records, unknown services, unexpected duplicates or unreconciled figures stop generation.

HumanHUMAN APPROVAL

7. Human review and send

Eloise reviews the generated report, confirms the figures and narrative are appropriate, and approves it before external release.

Human control

The pilot retains human approval before any client report is sent. Any discrepancy is corrected before release.

Report source map · draft alternative widget

Where the current report data comes from

The report stays visually intact. Source pills on the right point back to the relevant report lines.

FUTURE CULTURE + ENTROPICO

FY’25 - FOURTH QTR REPORT

APRIL 1 - JUNE 30

NUMBER OF CLIENTS USING SERVICE THIS QUARTERTwo

NUMBER OF SESSIONS COMPLETED IN TOTAL THIS QUARTEREight

TOTAL NUMBER OF EAP SESSIONS USED THIS CONTRACT YEARTwenty-nine (out of 40)

TOTAL NUMBER OF EAP SESSIONS BOOKED FOR NEXT QUARTEROne

REASONS FOR SESSION BOOKINGS• General Mental Health Support
• Workload/life Balance and Management

NOTES & RECOMMENDATIONWe acknowledge another quarter of EAP service for Entropico, with two people using the EAP service between April and June booking eight sessions in total. This means a total number of twenty-nine out of forty sessions have been used in total. We are also very proud to share with you our Elavet Leadership Lab program brochure. It is an honour and privilege to be supporting Entropico’s workplace in this way.

Source: Zanda Sales report
→ Number of clients using service

Source: Zanda Sales report
→ Sessions completed this quarter

Source: Zanda Sales report
→ Cumulative sessions

Source: Reporting Register
→ Annual allowance

Source: Zanda Sales report
→ Sessions booked next quarter

⚠ Source to confirm with Eloise
→ Reasons for session bookings

Source: validated report facts + approved context
→ Notes & Recommendation

Remaining feasibility and testing

What Empire Ops still needs to review

1	Status-filter extraction: after Eloise confirms which Appointment Status values count against the allocation, confirm the simplest repeatable way to export those records from Zanda Sales. Because status is a Zanda report filter but is not shown as an exported column, the process needs a clear way to identify which status-filtered export it is using.
2	Spreadsheet cleaning: take a representative exported Sales Excel file and, in the spreadsheet/processing step, filter the Type column to Service. Confirm this removes the associated Session Pack rows and leaves one Service row per appointment. This is a filter applied to the exported spreadsheet, not another Zanda report filter. Only if duplicate Service rows remain do we need the candidate Client + Practitioner + Appointment Date/Time deduplication rule.
3	Live Reporting Register: if Eloise approves the approach, test that the ChatGPT reporting workflow can reliably read the current Google Sheet, find one matching client/allocation, and stop if the entry is missing or ambiguous.
4	Report generation: test the approved Google Doc master for branding, variable content, validation and one-page pagination, including keeping the contact statement together.

Design Check closeout

Are we aligned on this future workflow?

If Empire Ops can complete the remaining review and prove the unresolved technical items work as described, is this the reporting workflow Future Culture wants to operate?

Zanda-first practitioner process: all reportable practitioners use Zanda consistently, including Virginia if she is delivering reportable organisational sessions.
Appointment data discipline: every appointment has the correct Service, Session Pack and final Appointment Status, with no past appointments left “Pending”.
Usage-status rule: Future Culture confirms which final statuses consume the organisational allocation and therefore count in reporting.
Weekly human control: admin checks past appointments and reporting is blocked until unresolved records are corrected.
Repeatable reporting process: Future Culture is comfortable exporting Zanda report “Sales” and applying the agreed cleaning/calculation process.
Live Reporting Register: Future Culture is comfortable maintaining the agreed client reporting setup and annual allowance in a simple live register.
ChatGPT report generation: the cleaned dataset is imported into the controlled ChatGPT reporting template, using structured fields such as client and reporting period.
Human approval remains: Eloise reviews and approves the generated report before external release during the pilot.
Remaining content decision: confirm the source/process for reasons for session bookings and the definition of number of clients using the service.

Later opportunities, not required for the initial pilot

Pre-built reporting workbook / dashboard

A later spreadsheet could contain the Reporting Register, raw Zanda export, deterministic processing and a QA dashboard.

Potential future Zanda API integration

Zanda's API is not currently publicly available. As a later optimisation, investigate whether future API access can provide the required reporting fields and reduce or remove the manual “Sales” export step. This is not required for the initial pilot and must be technically verified before being treated as feasible.