Irana Wasti, CEO of Thought Industries, argues that B2B AI initiatives fail on the knowledge they’re built on, not the model, and that connecting fragmented systems is not the same as unifying them. She and Greg Kihlström get into what a real customer knowledge architecture requires.
A chatbot pointed at existing docs inherits every contradiction in them. Marketing, sales, support, and product each hold their own version of the truth, and automating on top of that delivers the wrong answer faster and in more channels.
The ROI case runs through the knowledge foundation, not the platform. Wasti on drawing a line from a unified knowledge layer to measures a CFO will accept, beyond case deflection and faster response times.
Acting is a different problem from answering. Once AI upgrades a plan or processes a return, the system needs context about the customer, the moment, and what it’s permitted to do, plus guardrails on which actions stay with a person.
Hosted on Acast. See acast.com/privacy for more information.
