Alteryx CEO: Why AI Agents Need Business Rules
22m
AI agents promise to close the books, price deals and run processes on their own. At work, however, a reasonable guess is not good enough. Alan Shimel talks with Andy MacMillan, CEO of Alteryx, about why AI agents need business rules, not just more context, before companies can trust them.
From Java developer to four-time CEO
MacMillan started as an enterprise Java developer helping General Motors build its first websites. He later held product roles at Oracle, ran Data.com at Salesforce and took UserTesting public. He joined Alteryx because the AI projects he built kept needing business context. Alteryx lets business teams in finance, supply chain and rev ops apply Python, SQL and machine learning to their data without writing code.
Why AI agents need business rules
Personal AI feels like magic because it fills in assumptions for you. Those same assumptions become a problem at work. An agent asked to calculate revenue recognition or a sales commission may give a different, reasonable answer each time. None of them may match company policy. Therefore, MacMillan argues, agents need business rules that are visible, understandable, repeatable and auditable before teams can trust them.
Business ownership and real ROI
Business rules change when tax law, pricing or a sales spiff changes, so business teams, not just developers, must own that logic. In addition, reusing trusted workflows costs far fewer tokens than asking an LLM to rediscover the same logic every time. That combination of trust, ownership and reuse is how MacMillan expects companies to see real ROI from agentic AI.
Explore more DevOps and AI videos and the latest Techstrong TV interviews.
For more information please visit the company website