The Artificial Intelligence Podcast

Kamal Hathi, SVP and GM of Splunk at Cisco, breaks down why intent is never enough to control an AI agent: agents are ruthless goal seekers, and asking one to free up 3 percent of disk space can end with 97 percent free because nobody said keep the rest. He explains why trust means verification rather than instruction, and why it includes the invoice, comparing it to hiring a contractor you trust to build a house that stands without emptying your bank account. Kamal also digs into why the rush toward air gapped and local AI is driven less by distrust of frontier labs than by cost, with specialized small models running 97 percent cheaper and 95 percent faster than frontier calls, and why the real threat is the unknown unknowns, agents nobody cataloged acting inside a state of perpetual breach. Finally, he argues agents are the new apps, and every job starts to look like a software developer's: tell the agent what to do, inspect the work, and iterate.

Be a guest on the show: https://tonyphoang.com/guest

Live interviews at your conference: https://tonyphoang.com/live

Music and credits: https://tonyphoang.com/credits

Hao Yang, VP and Head of AI at Splunk, breaks down why sampling an AI system's output cannot earn trust: a fraction of ten thousand probabilistic calls will be wrong, and a spot check is unlikely to find them. He explains why a small model can grade a frontier model reliably inside a narrow context but not outside one, and why prompt injection follows from the architecture rather than being a bug, since following instructions is how these systems work. Hao also digs into why time series is still not a first class citizen in frontier models even as multimodal training reaches vision and audio, and why he rejects routing prompts to the cheapest model, arguing cost is meaningless without value. Finally, he argues against the single super agent, which by design holds access to everything it might ever need, and for guided autonomy, where an agent asks before first reaching for a tool and the user grants it once, for the session, or never.

Be a guest on the show: https://tonyphoang.com/guest

Live interviews at your conference: https://tonyphoang.com/live

Music and credits: https://tonyphoang.com/credits

Mangesh Pimpalkhare, SVP and GM of Splunk Platform at Cisco, breaks down why agentic AI generates more telemetry than the systems it monitors, and why nobody can yet say when observing AI costs more than running it. He explains the decade spent preaching centralized data was always partial, since machine data never sat in one system, and that what broke was volume rather than architecture, pushing a whole category to abandon the ingestion pricing it ran on for 15 years. Mangesh also digs into why an agent acting on a bad signal is usually a context failure, not a data failure, since a model trained on everything still lacks up-to-the-second context, and why autonomy arrives on a sliding scale against governance as trust is earned. Finally, he argues control, not data, is the moat, since customers own their data and fragmentation only grows, leaving the advantage with whoever supplies the guardrails that stop agents doing the wrong thing.

Be a guest on the show: https://tonyphoang.com/guest

Live interviews at your conference: https://tonyphoang.com/live

Music and credits: https://tonyphoang.com/credits

Johan Land, Chief Product Officer at Samsara, breaks down why every major AI lab crowded into the easy 90% of knowledge work while ignoring the 40% of the economy that is physical and the 80% of workers who never sit at a desk. He explains why intelligence on a battery budget has to live on the edge - collision warnings fire in milliseconds, making GenAI's hundreds of milliseconds of latency unthinkable behind the wheel - while the heavy pattern-spotting waits for the backend. Johan also dives into why the era of dashboards is over, describing an agent that calls a drowsy driver over the in-cab camera to talk him into a coffee break, and why coverage across 99% of US roads beats a robotaxi fleet's high-resolution sensors. Finally, he argues frontline augmentation lasts indefinitely, since a delivery that navigates the dog and sets down the package is far messier than driving the van autonomously. This episode contains sponsored content.

Ketan Karkhanis, CEO of ThoughtSpot, breaks down why two decades of self-service BI was a hoax that only handed everyone the right to build their own dashboards, leaving companies sitting on 50,000 of them with no idea what to do. He explains why the industry's shortcut of bolting a language model onto your data and letting it write the SQL is doomed: analytics are deterministic while AI is probabilistic, so the same churn question comes back three different ways and a frontier model on a warehouse lands near 50 to 70% accuracy, a hit rate no other enterprise function could call a product. Ketan also digs into why legacy vendors are just repainting 20-year-old tech with AI, and why he runs agents that own a function rather than a task, like one that approves NDAs end to end and escalates only the 5% it can't. Finally, he shares why the analyst clinging to dashboards goes extinct while the one who becomes the "AI steward" of governed data wins, as data teams shift from serving humans to serving agents. This episode contains sponsored content.

TVN Reddy, CEO of Aptean, breaks down why the entire software industry is selling "bigger is better" AI while the real goal is the opposite: people spending 80% less time in systems and more time doing the work that actually drives the business. He explains why general purpose models hand every competitor the same 95% while the deep vertical 5% drives the outsized gains, why enterprise AI is unusable past roughly a half percent error rate, and why validation agents beat the instinct to hire three people to check the AI's work. TVN also digs into the quiet security crisis of employees pasting P&Ls and proprietary recipes into public LLMs that never had training turned off, and why agents bolted onto generic models won't scale without truly vertical models underneath. Finally, he shares why the winners won't be the companies with the hottest models but the ones with the courage to re-engineer their processes, because every big technology removes old constraints only to create new ones. This episode contains sponsored content.

Bill Franks, President of Analytics Advisory Partners, breaks down why a degree no longer proves much and employers now hire on what you can show, not what you know. He explains the pipeline trap companies are walking into - handing every junior, tactical task to AI the way firms offshored it back in 2000, then finding no one qualified to step into the senior roles - and argues teams should deliberately do one in ten tasks by hand just to keep their skills sharp. Bill also digs into the boring physical limits that could throttle AI long before the algorithms do, from power and water to grids so backed up that wiring a data center now takes longer than building it. Finally, he unpacks model collapse, where models trained on AI output slowly degrade until the only fix may be retraining on human content from 2022 or earlier. This episode contains sponsored content.

Vishnu Hari, CEO of Ego AI, breaks down why the labs with billions in compute missed the product hiding in plain sight: character AI quietly tops every consumer retention leaderboard while the serious labs write it off as a sideshow. He argues humanness itself is the hard problem - dangerous, uncontrollable, and unchained - so the giants retreat to safe coding agents and ship characters that collapse into therapy or role-play because they never actually live lives. Vishnu takes apart the memory paradox (an agent that remembers everything is unusable; one that remembers nothing is forgettable) and points to the Nemesis system in Shadow of Mordor, a feature so good Warner Bros trademarked it, as proof you can build real relationships without superintelligence. He also gets into the edges no one wants to touch: who inherits the AI you've talked to for a decade, and why the new "I'm not human" disclosure laws are pure theater. This episode contains sponsored content.

Carter Huffman, CTO of Modulate, breaks down why voice AI keeps failing in production despite years of "solved" transcription claims. He explains the "transcript trap" that fools dev teams into shipping voice agents that crumble on real calls, why bigger language models actually make latency worse, and how ensemble approaches with smaller specialized models outperform monolithic systems. Carter also dives into the explosion of deepfake voice fraud - including the $25M Hong Kong heist - and why passive monitoring plus red teaming are now essential to a modern voice security stack. Finally, he shares why even Gen Z still defaults to voice for high-stakes issues, and what contact centers must get right over the next three years to build genuine trust with callers. This episode contains sponsored content.

Join Babak Hodjat, Chief AI Officer at Cognizant, as he unpacks what's actually happening at the cutting edge of agentic AI — from agents that run continuously without being asked, to the governance crisis quietly unfolding inside enterprises right now. Babak draws on decades of AI research to break down multi-agent architecture, the TerraLingua experiment that let AI agents form their own societies, and a breakthrough in evolutionary fine-tuning that could change how the entire industry thinks about customizing large language models. This episode contains sponsored content.