From Pilots to Production: Agentic Commerce, Enterprise Trust, and the Quantum Horizon

In this episode of the Disambiguation podcast, host Michael Fauscette talks with Julia Samoylenko, Founder and CEO of Asteri AI, about why deploying AI agents without redesigning the workforce around them is a missed opportunity, how her platform embeds workforce redesign directly into the product rather than leaving it as a customer problem, and what it actually takes to orchestrate humans and AI agents as a governed team inside highly regulated life sciences organizations.Julia spent 11 years at Medtronic in cross-functional roles across multiple regions, most recently as Director of Transformation, before founding Asteri AI in Silicon Valley. The company builds what she calls an AI operating system for life sciences, a platform that enables enterprises to run humans and AI agents as one governed workforce. Asteri AI maps how work gets done across the enterprise, identifies where agents would perform best, and then orchestrates the resulting human-AI team with built-in governance, traceability, and continuous measurement.The conversation covers why deploying agents without understanding the full picture of work leads to fragmented results, how Asteri AI starts by mapping work across the enterprise before proposing which agents to deploy and where, why workforce redesign should be built into the product rather than treated as a separate consulting engagement, what a governed human-AI workforce looks like in practice with full traceability of every agent action and human approval points at every step, how the platform automatically proposes agent deployment based on where inefficiencies exist and where AI would have the most impact, why measurement should focus on capacity increase rather than cost savings alone, how agent work differs across regulatory submissions, clinical operations, and quality and safety, why life sciences leaders need to think beyond single agents and take a systems view of their entire function, and why change management for AI is a continuous process rather than a one-time event.Timestamps:00:00 - Introduction00:31 - Episode title and guest intro00:45 - Julia's background: 11 years at Medtronic01:19 - What led her to start Asteri AI03:06 - What "AI operating system for life sciences" means04:24 - Increasing capacity through human-AI orchestration05:02 - The holistic view: treating agents as team members05:52 - Company-wide deployment and the work intelligence module06:43 - Continuous value measurement as a differentiator07:57 - Shifting from technology focus to workflow focus08:28 - What a governed human-AI workforce looks like09:56 - Workforce redesign built into the product10:39 - Starting with how work gets done, not with the agents12:16 - Change management: still required but now data-driven13:27 - Deep vertical knowledge as competitive advantage13:44 - Governance and oversight in regulated industries15:40 - Scaling past pilots: capacity and measurement16:43 - Capacity increase vs. cost savings17:07 - Regulatory submissions: measuring throughput and time18:03 - Process simulation for decision makers18:58 - Growing capacity without proportional headcount growth19:32 - How enterprise experience shapes the product and earns trust22:13 - How AI work differs across clinical, regulatory, and quality25:11 - What life sciences leaders should know before deploying agents27:05 - Roles becoming obsolete and new work emerging28:37 - Workflow-first approach vs. technology-first29:00 - Change management as an ongoing process30:42 - Orchestration, governance, and workflow redesign as priorities31:01 - Recommendation: Thomas Otter and Ethan MollickGuest: Julia Samoylenko, Founder and CEO, Asteri AIHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

In this episode of the Disambiguation podcast, host Michael Fauscette talks with Daniel Fallmann, Founder and CEO of Mindbreeze, about why the most capable AI models in the world are still guessing without an enterprise knowledge layer, how the consumer of enterprise knowledge has shifted from humans searching for documents to AI agents performing tasks, and what it actually takes to make twenty years of enterprise information usable by the next generation of AI.Daniel founded Mindbreeze in 2005 at the age of 23, right after his computer science studies, and has spent two decades building the company into one of the leading enterprise search and knowledge management platforms. Mindbreeze is part of Fabasoft Group, headquartered in Linz, Austria, with a North American headquarters in Chicago and customers across Europe and the US. The platform connects to more than 500 enterprise data sources, understands permissions and context, and now serves as the knowledge layer for AI assistants, agents, and applications.The conversation covers why the consumer of enterprise knowledge has changed (it is no longer just a person searching for a document but an AI assistant, an agent, or an application that needs context in real time), what the enterprise knowledge layer needs to look like before AI can work reliably (connectivity, permissions, semantic retrieval, and focused context rather than more context), why agents should not know everything but only what they need for a specific task and a specific user, how Mindbreeze InSight Touchpoints move organizations from search and retrieval to orchestrated workflows where agents select the right tools and get work done, why forward deployed engineers matter in enterprise AI (connecting SharePoint is a technology task but understanding which information matters for an engineering process and how it relates to SAP data is a business problem), why AI is displacing individual activities rather than entire professions (the question is not whether AI eliminates a role but which specific tasks no longer need to be done manually), why change management for AI should be continuous work redesign rather than a one-time event, what separates AI native organizations from AI augmented ones (adding a copilot to every application does not make a company AI native), and the three questions every C-level leader should ask before scaling AI.Timestamps:00:00 - Introduction00:45 - Daniel's background: founding Mindbreeze at 2304:04 - Mindbreeze as the knowledge interface for AI agents and applications05:09 - Why models are no longer the biggest problem enterprises have06:48 - Enterprise search as the knowledge layer for AI07:46 - Your AI will change but your enterprise knowledge will not08:34 - The strategic role of the knowledge layer12:16 - Forward deployed engineers and the last mile of enterprise AI14:41 - Turning customer knowledge into productive AI native use cases15:36 - Adoption, trust, and getting people to use AI18:40 - More context is not better context19:08 - Build the knowledge layer first, then scale agents on top20:30 - Why the bottom-up approach works better than top-down22:41 - Where AI is displacing knowledge work fastest23:11 - Displacement means activities, not entire professions25:03 - Start with which tasks consume thousands of hours today27:33 - Winners will continuously redesign how humans and AI divide work28:51 - Change management is a process, not an event29:14 - US vs Europe: speed versus governance30:15 - Speed without governance creates risk31:17 - Governance without execution creates no value33:28 - AI native versus AI augmented organizations34:02 - Adding a copilot does not make you AI native38:33 - Models become replaceable when knowledge is your foundation39:43 - Enterprise knowledge is becoming infrastructure40:27 - RecommendationGuest: Daniel Fallmann, Founder and CEO, MindbreezeHost: Michael Fauscette, CEO & Chief Analyst, Arion Research

In this episode of the Disambiguation podcast, host Michael Fauscette talks with Awais Rafeeq, Founder of AI Data House, about what actually breaks when companies try to move AI from pilot to production, why most small and mid-sized businesses are still running critical operations on disconnected spreadsheets, and the practical playbook for building automation that sticks.Awais started his automation career before the AI wave. As a mechanical engineering graduate working as a data analyst, he built his first automation using Excel VBA to evaluate e-commerce products, a process that used to take two to three days per product. He sold that tool to classmates for $100 each and never looked back. That side project evolved into data analytics, reporting dashboards, and RPA-based process automation before AI entered the picture. Today, AI Data House has completed more than 500 business automation projects, primarily for small and mid-sized businesses in the US.The conversation covers why companies are still running seven people on seven separate Excel files with no central system (and why the first step is not AI but centralized collaboration), the most practical starting points for automation (connecting Gmail leads to a CRM, adding chatbots, deploying calling agents for after-hours coverage), what breaks when a company tries to move from pilot to production (people hesitation, the plug-and-play myth, and the need for 15 to 20 days of monitored maturation), why you cannot automate a broken process (and what it takes to document processes, define gates, and map edge cases before any AI touches them), how to build trust with AI phase by phase (start a chatbot on FAQ only, add refund processing later, begin dashboards as read-only before enabling optimization), what healthcare AI taught him about compliance and data residency (you cannot give patient information to Gemini or Claude, everything stays in compliant containers), why small businesses move faster than enterprises (convince one or two people, get full context from one or two people, and start building), and the change management approach that actually works (automate one part of someone's job, show the results, and they will ask for more).Timestamps:00:00 - Introduction00:33 - Episode title and guest intro01:07 - Awais's background: mechanical engineering to automation01:47 - The first automation: Excel VBA for e-commerce product evaluation03:06 - Patterns in SMBs: still running on disconnected spreadsheets04:23 - Starting points: Gmail to CRM, chatbots, calling agents05:19 - What breaks moving from pilot to production05:35 - People hesitation and the plug-and-play myth06:19 - Monitoring, maturation, and the importance of logging accuracy07:20 - Data problems when walking into a new client08:33 - Documented processes and process ownership09:21 - Trust building: deploying chatbots and agents phase by phase10:59 - Read-only dashboards before enabling AI optimization11:46 - Healthcare AI: compliance, data residency, and compliant containers13:17 - Automating referral processing and calling agents in healthcare14:36 - Why small businesses move faster than enterprises16:09 - Change management and employee resistance16:58 - Fear declining: adopt AI or be replaced by someone who does18:29 - Automate one part, show results, they ask for more19:26 - Finding high-value starting points for automation22:01 - You cannot automate a broken process24:25 - RecommendationGuest: Awais Rafeeq, Founder, AI Data HouseHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

In this episode of the Disambiguation podcast, host Michael Fauscette talks with Vasanth Mohan, Product Marketing Lead at SambaNova Systems, about why inference speed is the bottleneck holding back AI agents, how purpose-built silicon changes the economics of enterprise AI, and what business leaders need to understand about the infrastructure layer that sits beneath every AI application they use.Vasanth's career has tracked the frontier of emerging technology. He studied computer science at Stanford, worked in VR during the Oculus era, moved into edge computing and telco networking, and landed in AI infrastructure as the bottleneck shifted from software to silicon. At SambaNova, he leads technical product marketing and developer relations for the company's reconfigurable dataflow unit (RDU), a chip designed specifically for AI inference.The conversation covers which AI use cases are gaining real enterprise traction (coding is the base layer, coworker tools are actually coding under the hood, and the world is moving through three waves from chat to RAG to agents), what premium inference actually means for business leaders (not just faster tokens but the sweet spot of running large models fast enough for agents to work), why a coding agent that takes 25 hours could run in under one hour with the right infrastructure (and why that changes everything about productivity), how disaggregated inference splits the pipeline so that GPUs handle one phase while purpose-built chips handle another (prefill is compute-bound, decode is memory-bound, and they need different architectures), what the dataflow architecture does differently from GPUs (operator fusion that streams data through operations instead of bouncing back to memory), why speed is now priced as a distinct tier (and when paying double for fast tokens makes sense versus batching overnight), why JP Morgan Chase and sovereign AI deployments across Europe and Asia Pacific are choosing on-premises inference (regulatory requirements, audit concerns about closed providers, and the fact that 80% of data centers are air-cooled), what developers are building on SambaNova Cloud (planning agents on frontier models delegating to fast open source execution agents), and the one metric enterprise leaders should use to evaluate inference infrastructure (tokens per second per megawatt, normalized for the service levels their applications actually need).Timestamps:00:00 - Introduction00:32 - Episode title and guest intro00:49 - Vasanth's background: Stanford, VR, edge computing to AI infrastructure05:00 - AI agent use cases gaining real traction05:27 - Coding as the base use case for enterprise AI06:20 - Coworker tools are actually coding under the hood07:00 - Three waves: chat era, RAG deployments, agent era09:18 - Premium inference: a distinct product tier, not just speed10:18 - Model size versus speed: the competing forces11:52 - Prompt caching is solved; decode speed is the unsolved bottleneck12:14 - The 25-hour coding agent: what 20x faster inference changes13:57 - Two components of agent workloads: CPU execution and inference15:14 - Disaggregated inference: splitting prefill from decode18:30 - Dataflow architecture and operator fusion on the RDU20:09 - Speed priced as a tier: when paying double makes sense22:41 - Voice agents need sub-500 millisecond response time25:00 - On-prem inference: JP Morgan Chase and regulatory requirements26:17 - Open source models are essential for secure environments27:36 - Sovereign AI and geopolitical data concerns31:58 - Right-sizing models to the task34:00 - Developer patterns: planning agents plus fast execution agents35:44 - Evaluating infrastructure: tokens per second per megawatt37:06 - Six-month ROI for on-prem NeoCloud customers38:08 - Recommendation: Semi Analysis and Dylan PatelGuest: Vasanth Mohan, Product Marketing Lead, SambaNova SystemsHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe so you never miss an episode.

In this episode of the Disambiguation podcast, host Michael Fauscette talks with Zoher Karu, Head of AI at Taelor, about why data governance has suddenly become a boardroom priority, how dirty data and missing business context cause AI projects to fail, and what practical steps leaders can take to build a data foundation that actually supports AI at scale.Zoher has spent his career at the intersection of data, analytics, and business strategy. He holds a PhD in electrical engineering from MIT, started at McKinsey, then founded startups in retail analytics and call center intelligence before leading enterprise-wide data and analytics organizations at Sears Holdings, eBay, Citibank (across 17 markets in Asia and Europe), and Blue Shield of California.The conversation covers why data governance is "sexy again" (AI amplifies data quality problems, so bad data now means bad decisions at machine speed), the blood-in-the-body analogy for enterprise data (every organ needs it, it should not be dirty or leaking), why a two-year data cleanup project is the wrong approach (clean as you go with a use-case-driven mindset), three reasons AI projects fail to deliver results (data quality and trust, missing business context that lives in people's heads not databases, and change management resistance), the "Susie knows how to do that" problem (business rules and institutional knowledge that AI agents cannot access), why pilot success does not predict production success (manually cleaned spreadsheets do not reflect real-world data), change management and the value exchange (people need to know what is in it for them), why productivity gains are not the same as business transformation (the real power of AI is reimagining processes entirely), cross-industry patterns in data challenges (siloed customer views, fractured definitions, departmental selfishness), how data teams are evolving toward full-stack roles with AI-assisted tools, governance as brakes that help you go faster (knowing the boundaries lets you push all the way to them), and practical first steps for business leaders (start with cost savings, ask employees what would make their job easier, build momentum through small wins).Timestamps:00:00 - Introduction00:32 - Episode title and guest intro00:45 - Zoher's background: MIT, McKinsey, startups, Sears, eBay, Citibank, Blue Shield03:18 - Business-first mindset for data leadership05:03 - Why data governance is sexy again06:18 - Data is like blood: the enterprise body analogy07:33 - When "revenue" means different things to different teams08:02 - Clean as you go, not a two-year cleanup project09:30 - The shift from AI productivity to AI governance10:00 - Three reasons AI projects fail10:55 - Missing business context: the "Susie knows" problem13:10 - Pilot versus production: the cleaned spreadsheet trap13:38 - Change management and the value exchange15:07 - Making existing work easier: the call center notes example15:28 - The real power of AI: reimagining processes entirely17:34 - Board-driven AI mandates and the fear of being left behind18:44 - AI is a tool, not a solution looking for a problem19:32 - Cross-industry patterns in siloed customer data21:49 - Citibank example: loan data missing time of day23:40 - Full-stack data teams and AI-assisted tools26:17 - Building governance that protects without becoming a bottleneck27:07 - Traffic laws analogy: rules of the road for AI29:32 - Brakes help you go faster30:14 - Practical first steps: start with cost savings and productivity32:06 - Ask your employees what would make their job easier33:04 - Success builds on success33:31 - Recommendation: Factfulness by Hans RoslingGuest: Zoher Karu, Head of AI, TaelorHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.