How We Agent

How We Agent

By Lyzr AITechnology
Download on the App Store

All How We Agent episodes

  • How the US DOT Used AI Agents to Analyze 120,000 Public Comments

    How do you bring AI agents into the largest organization in the world?

    In this episode of How We Agent, Ani sits down with Pavan, former Chief Digital and Information Officer at the US Department of Transportation and former CTO of the Federal Motor Carrier Safety Administration (FMCSA). Before government, Pavan spent nearly 20 years in the private sector, including leading customer-facing tech for Walmart International. He has also taught at Rochester Institute of Technology and studied at Columbia.

    Pavan shares how the DOT used AI agents to analyze tens of thousands of public comments on proposed rules, modernized 20-year-old legacy systems nobody could maintain, and structured AI teams, governance and disclosure inside a federal agency. He also explains why government should be the role model for digital experience, not the laggard.

    The conversation covers agentic AI in government, data sovereignty, legacy modernization, public datasets anyone can build on, measuring AI by total cost of ownership instead of tokens, and why so few AI agents make it to production.


    43 min
  • How Department of Growth Turned 17,500 Pages Into an AI Hologram

    What if you could have one more phone call with someone you’ve lost?
    In this episode of How We Agent, Ani sits down with Brian Dema, founder of Department of Growth, a fractional marketing agency for fast-growing startups. Five years after his father, J. Richard Dema, passed away, Brian and a 20-person team spent six months turning 17,500 pages of his writing, client work and archives into RichieBot. RichieBot is an AI version of his father with a cloned voice, an Unreal Engine face and a 4-foot interactive hologram.

    Brian walks through how they built it, from scanning five filing cabinets and building a RAG system to writing “soul files” on top of Claude. He then explains how they tested it with a Stanford psychologist to measure how close the bot came to the real man. The answer surprised them.

    This is a story about AI, and also about family, grief, memory and what we want to preserve of the people we love.

    52 min
  • A Breakdown of What Cisco’s AI Infra Stack Looks Like

    In this episode of How We Agent, Ani sits down with Raj Rikhy, Senior Director of Product Management for AI Platform at Cisco. Raj has spent nearly 25 years at Microsoft, IBM, Salesforce and Cisco, with half of his career in infrastructure and half in AI, including Microsoft’s early responsible AI work.

    Raj explains why Cisco is building a fully integrated AI stack, from Silicon One chips and networking to AI Defense, open-weight models and Splunk’s agent builder. He also explains why agent traffic is putting pressure on infrastructure and why the operational side of agents is underappreciated.

    The conversation covers AI infrastructure, agent security and guardrails, small open-weight models, personal agents for every employee, Cisco’s AI acquisitions, the forward-deployed engineer model, and why it’s a great time to start a career in product management.
    In this episode you’ll learn:

    • Why AI infrastructure is about much more than GPUs
    • How agents are sending traffic across networks at an exponential rate
    • Why almost 50% of IT professionals already use agentic capabilities
    • What Cisco AI Defense does: guardrails, harms mitigation and red teaming tailored to each agent
    • Why built-in cloud guardrails aren’t enough for production agents
    • Why smaller, fine-tuned open-weight models are winning for narrow, high-value tasks
    • What Cisco learned from giving every employee a personal agent
    • How Cisco’s acquisitions fit into an end-to-end agent portfolio
    • The skills new grads need to succeed as product managers in the age of agents

    Whether you’re building AI infrastructure, securing agents in production or starting a career in product, this episode gives you an inside look at how one of the world’s biggest infrastructure companies is approaching agentic AI.

    32 min
  • How DXKulture Built Market Research Agents to Scale Product Launches

    n this episode of How We Agent, Ani sits down with Sandy, founder and owner of DXagency and DXculture. Sandy has spent more than three decades in marketing and advertising, including roles at Lifetime, Fuse TV, Nickelodeon and Sci Fi Channel, and has run DX since 2009.

    Sandy explains how her team built a synthetic audience platform with Lyzr. It’s a closed-loop system of AI agents that act as real-time focus groups, combining authentic data with social sentiment to predict how real consumers will react.

    The conversation covers multicultural marketing, why general-purpose LLMs fall short for audience research, synthetic focus groups, product and pricing testing, getting access to first-party data, and why curiosity matters more than ever for founders.

    Whether you’re a marketer, an agency leader or a founder building with AI agents, this episode shows how synthetic audiences are changing the way brands understand their customers.

    30 min
  • How FedEx Is Using Agents to Automate Shipping Workflows

    In this episode of How We Agent, Ani sits down with Sri Kandikonda, Group Product Manager at FedEx, who leads the evolution of FedEx’s API ecosystem across 200+ countries. Sri has 15+ years in e-commerce, logistics and fintech, teaches a Maven course on taking AI agents to production, and wrote a book on the API economy.

    Sri shows how his team of product managers works alongside AI agents every day. He shares an internal agent that resolves 80% of product questions and explains why “anyone can build an agent, but few can build a reliable one.”

    The conversation covers agent guardrails, evals, governance, LLM selection, token budgets, building internal before external, and why AI agents need real-time data to make decisions.

    42 min
  • How EWN Is Using Agents in Its SDLC

    In this episode of How We Agent, Ani sits down with Jason Knight, who leads Agentic Transformation at EWN, a Texas-based operator qualification platform for the oil and gas industry. Jason moved into the role from inside the delivery organization. He has a background in Scrum coaching, Agile development and Lean Startup, with earlier stints at CommunityCare, BOK Financial and the Agile Leadership Institute.

    Jason shares how EWN went from running one or two experimental backlog items per sprint to rolling out agents across the company. He also explains why rising AI spend is a sign of progress and why human judgment matters more now than it did a year ago.

    The conversation covers agent orchestration, the BMAD Method, agent scorecards, governance frameworks, the future of Scrum, and what AI means for people who have spent years honing their craft.

    Whether you’re leading AI adoption, running agile teams or figuring out where you fit in an AI-first workplace, this episode is a practical look at agentic transformation as it’s happening.

    34 min
  • HFS Research: What the 5% of AI Agents That Make It to Production Do Differently

    In this episode of How We Agent, Ani sits down with Joel Martin, Chief Research Officer at HFS Research. Joel has spent nearly three decades in tech, with stints at IDC, Microsoft and TechInsights, and has led teams across North America, India, the UK and the Middle East.

    Joel explains how HFS went from stacks of PDFs to a live, queryable knowledge engine that customers can plug straight into their own systems. He also shares why subscription-based research is dead and what founders and enterprise leaders should focus on as AI changes every three months.

    The conversation covers enterprise AI adoption, the AI velocity gap, services as software, the headless future of work, AI risk and sovereignty, and the 4 Ps framework for getting AI into production.

    Whether you’re building an AI startup, leading enterprise transformation or managing teams through the AI shift, this episode gives you a research-backed view of where enterprise AI is heading.

    44 min
  • LionAIs: How a Cardiologist Used 16 Agents to Boost Case Reviews by 60%

    Dr. Lucia Jewbali, with 15 years in the ICU, frames her tool the way she frames the stethoscope: an instrument that augments one specific clinical sense and in this case, reasoning at the bedside.

    LionAIs, built with Lyzr, connects to the hospital EHR and combines symptoms with EKGs and bloodwork the way a clinician actually would.

    The design choice that matters is the constraint. Narrow the corpus, cite the source, and the tool becomes something a specialist will put next to a patient.

    30 min
  • How KPMG Takes Agents Through the Last 40%, from Demo to Production

    Swami Chandrasekaran has worked through four generations of AI and leads KPMG's move from experimentation into scaling. To understand agents properly, he built an agent harness from scratch rather than reading about frameworks. His mandate splits three ways: innovate, enable, standardize. Rather than build everything, KPMG assembles an ecosystem, which is why he is unusually direct about what gets a startup through the door: be a journey partner, bring differentiated technology, and have the nerve to let an enterprise actually use the product. The through line is the last 40%. Security, trust, scale and adoption are unglamorous, and they are what decides whether an agent ever leaves the demo environment.

    23 min
  • How Hitachi Put 200+ AI Agents into Production Across 280,000 Employees

    Patrick Corcoran leads marketing and external relations at Hitachi Digital Services, and his starting point is that AI is not new here. Machine learning and computer vision have been running inside Hitachi's factories and rail networks for around 25 years. What that bought them is a shorter run-up to the agent wave. HARC Agents is the enterprise framework: pre-built agents, unified observability and centralized control, with 200+ agents already in production and 618 sister companies in a position to test them. The part worth borrowing is where he stops. In the Lisbon rail command center, agents assist and a person still makes the call. Governance, security and observability are treated as the product, not the paperwork.

    27 min