Scheduling
Agent

The Context

Iro AI was selected for the Build3 startup accelerator cohort in May - July 2025. On pitch day, the Build3 team saw the system direction and asked us to build this automation for a workflow they were struggling with themselves.

That made the opportunity unusually strong: we were not pitching an abstract idea. We were stepping into a real operational bottleneck that the client already felt every cohort.

The Problem

Founders applied to Build3's fundraising program through the official website. After basic validation, applications flowed into Zoho CRM, where mentors reviewed and qualified each one manually.

Once a founder was qualified, the next step was scheduling an interview with one of four mentors. In practice, that meant someone on the team calling founders one by one, checking mentor calendars, confirming availability, and updating CRM records by hand. It was repetitive work, but it still demanded real coordination accuracy and real time.

The Solution

We built a fully automated scheduling pipeline that begins the moment a lead is marked as qualified in Zoho CRM.

The team only stepped in when a record needed manual correction and the flow had to be rerun. Otherwise, the system owned the full workflow from trigger to confirmation.

System Architecture

The architecture was designed as a full operational loop, not a standalone call bot. Every state change mattered: CRM qualification, live slot checking, mentor assignment, invite creation, transcript logging, and failure handling all had to stay in sync.

Architecture diagram for the Scheduling Agent system used in the Build3 pilot.

The mentor-matching logic accounted for availability, load distribution, and fit across four Build3 mentors. For calls that did not end in a completed schedule, the failure reason was written back into CRM, the record was corrected manually, and the system was rerun until completion.

Build Snapshot

Pilot Results

Thirty-three of forty founders were scheduled in a single automated call, with no follow-up and no human intervention. The remaining seven required a manual CRM correction and a rerun of the flow. Every one of them was subsequently scheduled, resulting in 100 percent final completion.

The broader outcome was bandwidth. Work that had consumed hours every cohort was converted into an automated process that ran while the team focused elsewhere.

Client Outcome

The relationship started with Build3 as accelerator and Iro AI as founder. It ended with Build3 as client and Iro AI as product team, which made this project as meaningful commercially as it was operationally.

Product Skills Demonstrated