Disambiguation

Michael Fauscette

"Disambiguation is the process of removing confusion around terms that express more than one meaning and can lead to different interpretations of the same string of text." Host Michael Fauscette of Arion Research; a leading technology analyst, tech startup advisor, consultant, board member, and storyteller; and his guests "remove the confusion around" artificial intelligence (AI), generative AI and business automation by looking at the business solutions available today to improve business outcomes and gain competitive advantage. 

  1. When AI Does the Building: Innovation, Ideation, and the New Creative Advantage

    3d ago

    When AI Does the Building: Innovation, Ideation, and the New Creative Advantage

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Dr. Alex Mehr, Founder and CEO of Famous Labs, about why the most important competitive advantage in the AI era is no longer engineering skill but taste, judgment, and knowing what to build. Alex argues that AI has made execution so much easier that the bottleneck has moved upstream: the people who will win are the ones with the strongest product instincts and the clearest sense of what the market actually needs.Alex's path is distinctive. He grew up in an academic family with plans to become a physicist and university professor. He earned a PhD in mechanical engineering and worked at NASA Ames Research Center in California. Through proximity to Silicon Valley, he caught the entrepreneurship bug and co-founded Zoosk, a dating platform that grew to include a large engineering team. He describes becoming an entrepreneur as crossing the Rubicon: once you do it, there is no going back. He even kept publishing research papers during Zoosk's early years just in case he wanted to return to academia. He never did. Now he runs Famous Labs, which he describes as the ultimate playground for really smart people, constantly launching new AI-powered products.The conversation covers the taste shift (why engineering thinking still matters but judgment and product instinct now move the needle more than raw coding ability), how Famous Labs hires (they no longer ask technical questions but instead ask "what have you built and why does it look that way?"), the junior engineer pipeline gap (real but short-term and already dissolving as younger engineers pick up AI tools), why layoff narratives are overblown (companies have always right-sized and AI is just the latest excuse), Famous Labs' multi-product model and why AI enables "idea machines" who can pursue multiple products because execution costs have dropped, the death of the software moat (SaaS companies can no longer rely on their code as a competitive advantage), human-centric product philosophy (building things that add value without taking value from other humans), the innovation process (every major Famous Labs breakthrough has come from getting smart people together in a literal hotel room), Heisenberg as "Cursor for chemists" (vertical AI for small molecule drug discovery using chemistry-specific foundation models), why vertical AI is the next major evolution (every profession needs its own cursor equivalent), the SaaS-pocalypse pushback (nobody is going to vibe code their ERP because the real moats are compliance, testing, integration, and business logic), the hybrid workforce concept (every knowledge worker must be AI-enabled or competitors will eat your lunch), game theory dynamics (AI lets competitors enter your territory the way calorie-dense potatoes enabled New Zealand's territorial unification), and practical strategic and tactical advice for executives.Timestamps:03:20 - The taste shift: judgment matters more than coding ability05:15 - How hiring has changed: "What have you built?" replaces technical interviews06:14 - The junior engineer pipeline and the education system10:05 - Layoff narratives are overblown: AI is just the latest excuse to right-size11:55 - The Industrial Revolution parallel15:11 - The software moat is gone15:54 - Human-centric products16:30 - AI as coworker17:55 - Context windows and training19:33 - The ideation to output pipeline20:16 - Innovation workflow25:16 - Vertical AI: every profession needs its own cursor equivalent27:51 - Specialized models for specific professions30:36 - The SaaS-pocalypse pushback: nobody is vibe coding their ERP34:25 - SaaS companies should use AI to make their tools 10x better37:32 - The hybrid workforce38:44 - Every knowledge worker must be AI-enabled43:11 - The New Zealand potatoes analogy: AI enables territorial expansion45:04 - Systematically enable each role with AI46:12 - Recommendation: Nassim Taleb and Antifragile

    48 min
  2. The End of One Model to Rule Them All: Why Enterprise AI Is Going Small, Specialized, and Multi-Model

    Jun 24

    The End of One Model to Rule Them All: Why Enterprise AI Is Going Small, Specialized, and Multi-Model

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Calvin Cooper, Co-Founder and COO of Neurometric AI, about why the dominant narrative of scaling ever-larger frontier models is giving way to a more practical reality: smaller, specialized models fine-tuned for specific tasks that are faster, cheaper, and more accurate for the vast majority of enterprise AI workloads.Calvin started his career in early-stage venture capital at NCT Ventures in the Midwest, then founded Rove, a consumer fintech company he took public via a Nasdaq direct listing. Now he and Rob May have co-founded Neurometric AI, which builds task-specific small language model infrastructure. They went full time in August 2025, at a time when the dominant narrative was still "scale compute, scale larger models, AGI," because they were seeing something very different in the research and in practical enterprise deployments.The conversation covers the surgeon analogy (why you do not hire a surgeon to schedule an email), how their leaderboard proved that no single model is universally best and that inference time tactics can be as impactful as model choice, the AT&T case study (scaling from 8 billion to 27 billion tokens per day while cutting costs by 90%), how 24/7 AI agent runtimes turned subscription costs into six-figure monthly inference bills, why 75% of enterprise AI tasks do not need a frontier model, their marketplace of 115+ task-specific models under 20 billion parameters with fixed monthly pricing per endpoint, the Coding Swarm (orchestrating task-specific SLMs across the development lifecycle), why AI coding agents prove that AI expands jobs rather than replacing them, the four-stage enterprise AI maturity model, why calling a bubble is intellectually lazy (railroads had a bubble too), GPU underutilization and the case for both scaling capacity and improving efficiency, edge compute as the next frontier, and practical advice for enterprises on multi-model orchestration.Timestamps:00:00 - Introduction00:44 - Calvin's background: VC at NCT Ventures, founding Rove, Nasdaq exit01:37 - Following curiosity: why inference is the largest market opportunity of our lifetime03:47 - The surgeon analogy: why frontier models are overkill for most tasks04:58 - Smaller specialized models are faster, cheaper, and more accurate06:03 - Ship fast: the leaderboard as first proof point06:26 - No universal good model: different models perform differently at different tasks07:26 - Early adopter customers and the enterprise journey07:57 - Real example: Llama model at 4x cost and latency improvement10:20 - AT&T: 8 billion to 27 billion tokens per day, cut costs 90%11:30 - The 24/7 agent runtime problem: from subscription to $100K/month bills13:09 - Multi-model orchestration as the natural next step14:05 - SaaS pricing disruption and the need for cost predictability14:53 - 115+ task-specific models under 20 billion parameters15:06 - Fixed monthly pricing per endpoint with frontier fallback18:01 - 75% of enterprise tasks do not need a frontier model18:57 - The Coding Swarm: task-specific SLMs for the development lifecycle20:34 - AI and jobs: coding agents expanded demand for developers23:09 - Stage 4 maturity: from monolithic AI to dynamic resource matching23:31 - First KPI is learning, not ROI28:16 - Infrastructure: existing GPUs are underutilized31:14 - Efficiency is not just cost: latency, privacy, compliance32:11 - Privacy and compliance reasons for multi-model architecture33:09 - No one God model: the future is less Mission Impossible, more Tron34:17 - VC perspective shaping the Neurometric business model37:08 - Practical advice: cut your inference bill by 80-90%39:28 - Wrap-upGuest: Calvin Cooper, Co-Founder & COO, Neurometric AIHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

    42 min

  3. Jun 17

    AI Meets the Mid-Market: How PE-Backed Companies Are Leapfrogging with AI

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Andrew Brooks, Founder and CEO of Contextualize, about why mid-market and PE-backed companies are in a unique position to leapfrog with AI, and how purpose-built solutions, inside-out disruption, and a multi-stage evolution from automation to intelligence are creating value these businesses could never have accessed before.Andrew is a serial entrepreneur whose career follows a consistent pattern: identifying new disruptive technology and connecting it to underserved markets. He founded SmartThings, the smart home platform that Samsung acquired, built and sold SMB Live to ReachLocal, and now runs Contextualize, which builds AI solutions specifically for mid-market B2B services organizations, many of them backed by private equity.The conversation covers why AI operates in two flavors (a new form of electricity and a tool for accelerating software creation), why mid-market companies now have the right to own purpose-built AI rather than renting features from enterprise vendors, how inside-out disruption differs from the Silicon Valley outside-in model, a fleet management case study where 14,000 emails per month from 3,000 vendors were processed by 13 humans, the vacation rental story, the multi-stage AI evolution from automation to data insight to prediction, "Digital Greg" and the challenge of capturing 25 years of institutional knowledge, governance by design with hard constraints, soft constraints, and separation of concerns architecture, how an agent layer can normalize data across 33 CRM systems after PE roll-ups, and practical advice for mid-market executives on where to start.Timestamps:00:00 - Introduction00:44 - Andrew's background: SmartThings, SMB Live, and founding Contextualize01:27 - The common thread: disruptive tech meets underserved markets03:08 - Why this is a leapfrog moment for the mid-market03:48 - AI in two flavors: new form of electricity and software accelerator04:43 - Own your AI, don't rent a feature05:06 - 25 years of institutional knowledge locked in people's brains05:40 - People, process, technology, and now AI as a fourth pillar06:24 - Inside-out disruption: how PE portfolio companies transform from within07:33 - Fleet management example: 14,000 emails, 3,000 vendors, 13 humans09:05 - The message to team members: removing tedium, not replacing people09:53 - 90% of solutions include a new human-AI interface10:49 - Vacation rental story: 3,000 properties, 10-12,000 work orders per month13:04 - The sidecar: a new human-AI interface for quality review13:50 - Ownership of outcome and the feedback loop14:18 - The batteries don't have serial numbers: edge cases that build trust15:25 - From checking to automating: the progression16:03 - Unexpected ROI: AI catches uninvoiced items16:48 - Multi-stage AI evolution: automation, then data insight, then prediction18:58 - Physical security company: hurricane-driven demand forecasting21:19 - Human in the loop vs. human in the lead at scale24:05 - You are never getting to 100%, and that is the right answer26:02 - Engineering firm: building code analysis with certification liability27:48 - Governance by design: hard constraints, soft constraints, and gating28:21 - Data governance as the most foundational layer31:07 - Don't over-index on security at the expense of value32:24 - Separation of concerns architecture with evaluator agents34:22 - Interceptor agents for cultural and behavioral guardrails36:33 - Digital Greg: capturing 25 years of refrigeration expertise39:42 - The line between AI and human touch is moving, not fixed40:44 - PE roll-ups and the 33-CRM nightmare41:26 - Agent layer for normalizing data across acquisitions46:08 - Advice for mid-market executives: where to start48:23 - Choose an internal champion49:33 - Recommendation: Thoreau's Walden, re-read at 51Guest: Andrew Brooks, Founder & CEO, ContextualizeHost: Michael Fauscette, CEO & Chief Analyst, Arion Research

    53 min

  4. Jun 10

    Beyond Efficiency: Why AI Is Forcing Marketing to Rethink Everything, Not Just Cut Costs

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Patrice Greene and Kathy Macchi, co-founders of Inverta, about why marketing's rush to AI efficiency missed the point, and what it really takes to rethink go-to-market workflows with AI at the core rather than bolted on top.Patrice is an early adopter of marketing automation who started in sports marketing before spending years in the Marketo community, eventually co-founding Inverta. Kathy brings an IT and operations background and has never had the luxury of separating marketing strategy from marketing infrastructure. Together they built Inverta to bridge the gap between strategy-led firms that lacked technical depth and tech-enabled firms that had no strategy, delivering what they call "roll up your sleeves" operational consulting for B2B marketing.The conversation covers what CMOs told Inverta's council at the end of 2025 (they thought they'd be further along with AI), why individual efficiency gains never translated into revenue impact, why you have to redesign workflows across teams rather than just hand out tools, the European supply chain analogy (why marketing needs its own ERP moment), the McKinsey threat (if marketers don't define how AI fits their function, consultants will define it for them), how CMOs need political capital and a vision that goes beyond cost cutting, Geoffrey Moore's four-box framework applied to AI decision-making in marketing, why managing AI agents has the same challenges as managing people (including a cautionary story about an agent that eroded a premium brand by over-optimizing for discounts), how AI is creating a new role in the buyer group and making 1-to-1 ABM at scale finally possible, and where marketing leaders should start their first AI workflow pilot.Timestamps (approximate, verify against final edit):00:00 - Introduction00:45 - Patrice and Kathy's backgrounds: from Marketo and IT ops to co-founding Inverta02:37 - Why Inverta exists: bridging the gap between strategy and tech in B2B marketing03:58 - CMO council findings: teams thought they'd be further along with AI05:07 - Individual efficiency gains did not translate into revenue06:22 - Don't leave adoption to chance: clarity, accountability, and support07:55 - Patrice: has efficiency really been realized? Now what?09:08 - FOMO is driving rapid adoption of AI point solutions09:56 - Automating broken processes just makes them broken faster11:29 - The European supply chain analogy: rethink the whole workflow13:17 - Who owns AI workflow redesign? Marketing, IT, or a translator?15:01 - Traction, not transformation: why the big word is counterproductive16:06 - Skills required: marketing expertise, org design, facilitation, change management16:55 - Marketing therapists: managing anxiety and fear in teams18:33 - Accountability: who is responsible when an AI workflow goes wrong?19:25 - The cost cutting trap: Gartner says you'll rehire 50-60% in two years20:40 - The story can't be all about efficiency: it has to be about growth21:17 - AI-mediated buyer journey: if you're not investing now, you won't even show up23:33 - The McKinsey threat: define AI's role or someone else will27:30 - Geoffrey Moore's four-box framework: core vs. context for AI decisions30:55 - Hybrid teams: workflow redesign before agents32:16 - Managing agents is like managing people: goals, guardrails, performance reviews33:29 - Brand risk story: agent over-optimized for discounts35:00 - Creating content for machines and people35:56 - AI as a new buyer group role: agents doing research on behalf of buyers37:30 - 1-to-1 ABM at scale: what used to be a luxury is now possible38:47 - Where to start: pick a workflow problem with a measurable outcome40:49 - Recommendations: Kerry Cunningham and Jeff WoodsGuest: Patrice Greene and Kathy Macchi, Co-Founders, InvertaHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode

    42 min
  5. The Cognitive Revolution in Leadership: Why AI Demands a New Human Operating Model

    Jun 3

    The Cognitive Revolution in Leadership: Why AI Demands a New Human Operating Model

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Victoria Mensch, CEO of Silicon Valley Executive Academy, about why AI is not just a technology shift but a cognitive revolution that challenges the very identity of leaders and demands a completely different human operating model.Victoria holds a PhD in psychology, spent 25 years in Silicon Valley high tech across large and small companies in enterprise software, and founded the Silicon Valley Executive Academy to help companies and executives tap into the Silicon Valley innovation playbook. Her unique lens, combining neuroscience, psychology, and leadership strategy, frames AI adoption as a human transformation challenge, not a technology deployment problem.The conversation covers why AI creates an identity crisis for leaders whose value was built on being the smartest person in the room, how the Silicon Valley innovation playbook applies to AI adoption (bias toward experimentation and treating failure as data), why human in the loop should evolve to human in the lead, the automation trap of applying AI to broken processes instead of redesigning work, why unrealistic productivity expectations are driving burnout, how AI unbundles job roles and creates both risk and opportunity, the shift from task management to systems design as the core leadership skill, and why empathy and motivation will define next-generation leadership.Timestamps:00:00 - Introduction00:45 - Victoria's path: PhD in psychology to 25 years in Silicon Valley tech02:19 - AI as a cognitive revolution: intelligence was the leader's identity03:39 - The identity crisis: machines can do cognitive tasks better04:11 - Finding your unique value: using AI as support, not replacement04:43 - The Silicon Valley innovation playbook: what the best companies do differently05:10 - Nobody has figured this out yet, even Silicon Valley is catching up05:39 - Bias toward experimentation: treating pilots as data-driven experiments06:34 - Embracing failure as a lesson, not a loss07:25 - From human in the loop to human in the lead08:05 - What leading an AI-augmented team actually looks like09:07 - What you put in is what you get out: the value of human input09:45 - Systems thinking versus task delegation10:07 - Managing AI teams is not that different from managing human teams10:50 - Subject matter expertise is not going away11:23 - Ownership mindset: "AI replaced my tasks" versus "I replaced those tasks"11:58 - Leadership versus position on the org chart12:30 - Treat your career as your business13:17 - AI unbundles job roles: what to automate and what to grow14:05 - Management versus leadership in the AI era15:04 - AI-accelerated burnout: the story of the marketing executive16:00 - The impossible expectation: performing at machine pace16:52 - Smart companies uplevel tasks instead of raising quotas17:20 - Burnout warning signs: chronic fatigue, lost motivation, physiological changes18:26 - Unrealistic productivity goals from executives who do not understand the tech18:44 - Do not outsource thinking: the value of cognitive work19:30 - Content flood: more output without more quality20:21 - Rethink the KPIs: what are you actually optimizing for?20:52 - Do not automate the broken process21:17 - Automating a patch that covers a workflow breakage just creates more noise22:19 - AI is a transformation opportunity, not just a tool23:10 - What it takes to redesign work at the organizational level24:59 - Three priorities: redesign work, build trust through clarity, elevate human qualities27:07 - The future of leadership: from task management to systems design28:54 - Empathic leadership and motivating free agents29:45 - Developer story: moving from coding to conceptual design30:49 - I want my engineers to solve problems, not write code31:47 - Recommendation: Sol Rashidi, CIO and AI thought leaderHost: Michael Fauscette, CEO & Chief Analyst, Arion Research

    33 min
  6. The Flight to Relationships: Why AI Is Making Trust the Ultimate Sales Advantage

    May 27

    The Flight to Relationships: Why AI Is Making Trust the Ultimate Sales Advantage

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Drew Sechrist, Co-founder and CEO of Connect the Dots AI, about why AI-generated outreach is flooding inboxes, destroying cold email effectiveness, and making trusted human relationships the most valuable asset in sales.Drew was employee number 36 at Salesforce, where he cold emailed Marc Benioff in 1999 and spent a decade helping take the company from zero to $1 billion in revenue. The biggest lesson from that experience: the cheat code in sales is knowing who knows who. Connect the Dots maps professional relationships using email history, LinkedIn career overlaps, and communication patterns, then scores relationship strength so sales teams can find warm paths into target accounts they never knew existed.The conversation covers Gresham's Law applied to outbound sales (bad outreach drives out good), why the only things that cut through inbox noise are trusted introductions and perfectly nailed problem statements, how the ghost email system works (the same approach Drew used with Benioff for a decade, now automated), why relationship strength should be a core primitive in every CRM system, the data quality challenge of building a 99%+ accurate relationship graph, the pendulum swing from data privacy fear to competitive FOMO, why AI native CRMs will challenge Salesforce and HubSpot, the barbell theory of future work, and why human relationships may be the last thing AI cannot automate.Timestamps:00:00 - Introduction00:42 - Employee 36 at Salesforce: cold emailing Marc Benioff in 199901:53 - The cheat code: it really is who you know03:38 - How Connect the Dots works: mapping invisible relationship paths05:12 - Finding warm paths you never knew existed: board members, college roommates, career overlaps05:53 - Proprietary scoring algorithm: relationship strength across your entire graph06:16 - The flight to relationships: Gresham's Law applied to outbound sales08:04 - The only two things that cut through inbox noise09:01 - Trust as the filter: if the messenger is trusted, you will read it10:18 - Ghost emails: how Drew turned Marc Benioff into his SDR for a decade12:04 - Automating the ghost email: reducing friction to one tap13:10 - The people with the most relationship leverage have the least time13:53 - How buyer behavior has shifted: 80% of buyers have already chosen their vendor15:30 - Relationship intelligence: planting seeds before buy mode begins16:54 - The economics of attention: trust earns the right to someone's finite time19:55 - Where agents should automate and where the human relationship stays20:48 - Tasks are going asymptotically toward zero, but relationships are the last holdout22:06 - The agent as presidential aide: facilitating, not replacing, the relationship24:17 - Data quality and privacy: three years to build a 99%+ accurate data engine25:13 - The pendulum swing: from data privacy fear to competitive FOMO27:33 - Not a data broker: intentional security and trust architecture29:42 - Where Connect the Dots fits in the evolving sales tech stack30:49 - AI native CRMs and the future of the CRM market32:21 - The trust layer across the internet: two new primitives for every CRM34:57 - 2026 is the year of actual AI automation of go-to-market workflows35:24 - Your relationship graph is the one proprietary signal your competitors cannot replicate38:57 - The hybrid workforce: the barbell theory of future work42:22 - The 10x engineer versus the 1.2x engineer44:47 - Recommendation: Bob Moore, CEO of CrossbeamGuest: Drew Sechrist, Co-founder and CEO, Connect the Dots AIHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

    47 min
  7. The AI Tax: Why Your Agents Cost More Than Your People and What That Means for Scale

    May 20

    The AI Tax: Why Your Agents Cost More Than Your People and What That Means for Scale

    In this episode of the Disambiguation podcast, host Michael Fauscette talks with Joshua Gould, CEO of The BigWord, about the hidden economics of enterprise AI deployment and why AI agents often cost more than the humans they are meant to augment.Joshua has spent over 20 years in language services, co-founded TBB Global, and now runs one of the world's largest language service providers operating in 80 countries across 250+ languages. The BigWord has been navigating AI disruption since the late 1990s, from machine learning-driven translation memory to neural machine translation to today's large language models.The conversation covers the real math behind AI agent deployment in call centers, why integration and infrastructure costs dwarf license fees, why the AI industry is negatively scaling at the macro level, the parallel between today's AI hype and the dot-com boom and bust, how regulated industries like courts and healthcare are deploying AI methodically versus recklessly, why governance by design matters when errors scale at machine speed, and why the companies built like cockroaches will outlast the hype cycle.Timestamps:00:00 - Introduction00:42 - Joshua's background: from selling beer to Wall Street language services03:40 - The first AI disruption: machine learning and translation memory in the 1990s05:23 - Integrations and automated workflows for banks06:54 - Pivoting to government contracting after the Great Recession08:25 - Building a defense contracting company from scratch to $20M10:49 - The 2019 vision: multilingual communications platform11:45 - Covid's devastating impact: losing 47% of revenue overnight12:46 - Selling to Susquehanna private equity in 202113:06 - LLMs arrive: the realization that AI is not free15:43 - The real math: why AI agents cost more than human agents18:07 - Hidden costs: integrations, infrastructure, tuning, and orchestration19:05 - AI can only do 70% of what a human does, 70% of the time19:54 - Why AI costs will not come down as fast as people expect21:10 - Data center rebuild cycles and the $1.4 trillion CapEx problem22:21 - Why deploy AI if it is not cheaper? Stability, speed, and service quality25:08 - The FOMO driving boards and the CEO firing wave30:03 - The coming correction: dot-com parallels and who will survive31:29 - Why enterprise SaaS is not going away32:19 - Staging AI deployment to protect against confidence collapse35:16 - Be a cockroach: companies built for survival versus hype37:01 - Governance by design: when errors happen 30,000 times in three milliseconds39:57 - Testing at scale and the danger of AI policing AI44:11 - Lessons from three waves of AI: watch regulated industries47:59 - Unintended consequences: data saturation and content noise48:40 - Recommendation: Gold with Gold podcast (Larry Gould, Cornell University)Guest: Joshua Gould, CEO, The BigWordHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

    51 min
  8. Governance Is Functions: Why Your AI Won't Scale Without Discipline by Design

    May 13

    Governance Is Functions: Why Your AI Won't Scale Without Discipline by Design

    In this episode of the Disambiguation podcast, host Michael Fauscette sits down with Chris Morancie, Fractional CTO and Founder of Digital Operations Factory, for a deeply technical and practical conversation about why AI governance has to be engineered into your architecture, not bolted on after the fact.Chris brings a unique combination of computer information systems, an MBA in business strategy, and a master's in data science to the problem of getting AI into production safely. His core argument: if your governance cannot stop your model from doing something wrong in real time, then it is not governance, it is just documentation.The conversation covers his three-part scalability test (design for scale, make sure it doesn't break at scale, don't go broke at scale), the Goldilocks zone for model selection, why agents should be treated through a microservices security lens with least-privilege access and short-term tokens, the firewall pattern for policy enforcement, real-time semantic interceptors for customer-facing AI, operational sovereignty and vendor SLA inheritance, IP leakage through model training, and a practical trust-vs-reasoning quadrant for managing hybrid human-agent teams.Timestamps:00:00 - Introduction00:44 - Chris's background: Caribbean upbringing, CIS + MBA + Data Science03:48 - The AI production framework: design for scale, don't break at scale, don't go broke at scale07:17 - The Goldilocks zone: model selection and cost benchmarking09:28 - Assertion testing vs. evaluation testing for model quality10:25 - "If your governance can't stop your model in real time, it's just documentation"13:26 - The firewall pattern: policy agents with least-privilege, short-term tokens16:09 - AI governance as good old-fashioned software hygiene17:49 - Real-time semantic interceptors for customer-facing agents21:15 - Competing goals: why prompts alone cannot prevent policy violations24:02 - Agent security: every ingress and egress point is a vector27:55 - RAG poisoning and downstream injection attacks29:00 - Operational sovereignty: SLA inheritance and vendor risk34:56 - IP leakage: when your feedback trains a competitor's model36:16 - Trust vs. reasoning: a quadrant for managing hybrid teams41:37 - Advice by company size: economics for SMEs, security for enterprise45:25 - Recommendation: DALI Research Labs (YouTube)Guest: Chris Morancie, Fractional CTO and Founder, Digital Operations FactoryHost: Michael Fauscette, CEO & Chief Analyst, Arion ResearchSubscribe and turn on notifications so you never miss an episode.

    48 min
See All (143)

Trailer

Ratings & Reviews

About

"Disambiguation is the process of removing confusion around terms that express more than one meaning and can lead to different interpretations of the same string of text." Host Michael Fauscette of Arion Research; a leading technology analyst, tech startup advisor, consultant, board member, and storyteller; and his guests "remove the confusion around" artificial intelligence (AI), generative AI and business automation by looking at the business solutions available today to improve business outcomes and gain competitive advantage. 

Information

  • Creator
    Michael Fauscette
  • Years Active
    2023 - 2026
  • Episodes
    143
  • Rating
    Clean
  • Copyright
    ©️All Rights Reserved
  • Show Website
    Disambiguation