In this episode, hosts Micah Smith and Kate Ressler spotlight winners from Automation Anywhere’s Agentic Bounty Challenge, created to help the Pathfinder community move beyond simple task automation and build goal-based AI agents for real business problems.
They interview three winners: Jenna Saunders, who built an agent that performs sentiment analysis on product reviews, outputs a CSV with sentiment scores, and generates actionable insights, using Python for precise data sorting and formatting; Omkar Mahajan, whose transit-focused agent uses maintenance and disaster management manuals to recommend next actions, create Salesforce tasks, work orders, and purchase orders with guardrails like order-history checks and approved product lists; and Ganesh Bhat, who built a “Sprint to Jira” agent that turns unstructured sprint planning inputs (screenshots, emails, transcripts) into Jira epics, stories, and tasks with human-in-the-loop and data protection guardrails.