The Challenge
What Ozzie Was Facing
Ozzie was an AI conversational assistant that needed to operate across three core business systems — a Salesforce CRM, a Zendesk support platform, and a proprietary knowledge base — without requiring agents to switch contexts. The systems used different authentication models, different data structures for customer records, and had no shared identity layer. Attempting to surface customer history during a conversation required manual lookups across all three, adding 3-4 minutes to every support interaction.
The Solution
What We Built
We built a systems integration middleware that created a unified customer context object by federating identity across all three platforms using email and phone as matching keys. A real-time context API allowed the conversational layer to pull current CRM status, open tickets, and relevant knowledge articles in a single request during session initialisation. Write-back connectors ensured that conversation outcomes were logged back to both Salesforce and Zendesk without agent intervention, eliminating duplicate data entry.
Results