We built a pipeline that moves student, attendance, and billing data from Brightwheel into BigQuery, then syncs the right contact records to GoHighLevel. Student and attendance data refresh daily; billing data refreshes monthly. Also, the client’s existing Loyalty Marketplace was connected to this new data source.
We standardized inconsistent Brightwheel fields, such as location names, room names, and charge categories. This gave the CRM and the Loyalty Marketplace a single, clean data structure to work with.
When the CRM couldn’t confidently match a contact, Claude helped select the best candidate based on name similarity and surrounding context. Clear matches still used deterministic rules; AI only handled the ambiguous cases.
Each run checks what changed since the last update and processes only new or updated records. This keeps the pipeline faster, cheaper, and easier to maintain as the data grows.
Sales call recordings are turned into useful insights: recurring questions, objections, coaching notes, and follow-up tasks. This helps the team learn from calls and maintain consistent follow-up.
The system reviews parent communication history, flags complaints or escalation signals, and scores churn risk. This helps the operator spot at-risk families before the issue becomes harder to fix.
The Slack bot answers staff questions using approved internal documents from Google Drive and Notion. If it can’t find the answer in the source material, it says so instead of guessing.
AI outputs are posted into Slack, so the team can review what the system found or decided. This keeps AI visible, checkable, and grounded in real source data.
