Web3 Unfiltered

As digital systems grow more complex, security has become one of the most critical and often misunderstood aspects of technology. For companies operating in cloud environments, fintech, healthcare, and emerging technologies like Web3, security is no longer just a technical challenge — it's a strategic requirement that directly affects trust, adoption, and long-term sustainability.

Lori Crooks is the founder and CEO of Cadre, a company focused on security assessments and security management. With a career dedicated to information security assessments, policy development, and compliance advisory, she has helped organizations navigate frameworks including NIST 800-53, FedRAMP, FISMA, ISO 27001, HIPAA, PCI DSS, and SOC audits.

In this episode, Collins Victory Odabi sits down with Lori to explore why companies underestimate security until it's expensive to fix, the real difference between being compliant and being secure, and how organizations should think about third-party risk and emerging technologies like AI.

In This Episode, We Discuss

• Security gets deprioritized because it isn't a profit center. Lori argues most organizations treat security as something that "won't happen to them," focusing budget on product growth instead — until a breach or compliance fine costs far more than building it in from the start would have.

• Choosing a framework starts with the type of data you hold, not the alphabet soup of options. Lori's practical guidance: FedRAMP for cloud systems selling to government, HIPAA for healthcare data, PCI for payment information — and the NIST Cybersecurity Framework as a strong general starting point for organizations without a specific mandate.

• Compliance can become a box-ticking exercise disconnected from real security culture. Lori describes organizations with unread policies and ignored training, stressing that most security failures are people-related — phishing links, shared passwords, gift-card scams — which makes genuine organizational buy-in more important than the paperwork itself.

• Vendor risk should be tiered by data sensitivity and access level, not treated uniformly. Lori recommends cataloguing every third party, assigning criticality based on what data or system access they hold, and reviewing high-risk vendors far more frequently than low-risk ones.

• Building security in from the start is cheaper than retrofitting it later. Lori cites CMMC's requirement for FedRAMP-equivalent tools as an example where companies that chose the wrong tools early face costly replacements later, while those who designed for compliance upfront can point to early third-party attestations as a competitive advantage.

About Lori CrooksLori Crooks is the founder and CEO of Cadre, specializing in security assessments, policy development, and compliance advisory across frameworks including NIST, FedRAMP, ISO 27001, HIPAA, and PCI DSS.

Key Timestamps / Chapter Markers[0:04] Introduction and episode framing[2:35] From accounting and IT auditing to founding Cadre[5:11] Why companies underestimate security until it's costly to fix[7:08] Choosing the right compliance framework to start with[9:16] The real difference between being compliant and being secure[11:34] Managing third-party and vendor security risk[16:19] What makes security training programs actually effective[20:38] Applying compliance frameworks to AI and Web3[24:26] Building security in from day one instead of retrofitting it

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

There's a growing assumption that AI is becoming intelligent — that models improve, systems become more reliable, more capable of decisions. That assumption is misleading, because AI doesn't understand anything. It predicts. And once probabilistic systems get placed inside environments that require precision and trust, something starts to break under pressure.

Jason Hart is a marketing data analytics and AI expert, multi-time SaaS founder, and analytics engineer who builds and audits the decision systems companies rely on. He's the founder of Domain Methods, working with SaaS and e-commerce companies on revenue data pipeline architecture and separating real AI value from noise.

In this episode, Collins Victory Odabi sits down with Jason to explore what AI is actually doing when it generates output, why monolithic prompts fail where atomic subtasks succeed, and why most companies' data simply isn't ready for AI at all.

In This Episode, We Discuss

• AI never gives you the same answer twice, even with identical inputs. Jason explains that generative AI works entirely through probability — predicting the next token based on likelihood, not certainty — meaning reliability requires testing an output not once, but repeatedly, to confirm it holds up.

• The biggest mistake is building one monolithic prompt instead of breaking work into atomic subtasks. Jason argues that stacking probabilities inside a single giant prompt compounds uncertainty, while decomposing a task into a sequence of small, controllable steps dramatically narrows what can actually go wrong.

• Deterministic business systems weren't built to tolerate a probabilistic component failing 1 time in 100. Jason describes the danger of dropping LLM-driven automation into unattended production pipelines that expect near-100% reliability, when large language models fundamentally cannot guarantee that.

• Most companies' data isn't ready for AI, no matter how good the model is. Jason routinely finds messy data hygiene, inconsistent metrics between departments, and almost no metadata describing what data actually represents — meaning an AI agent will confidently generate wrong answers based on wrong assumptions.

• The real fix isn't a smarter model — it's making the business itself as unambiguous as possible. Jason's practical answer: enforce shared definitions, documented processes, and a single reconciled source of truth across departments, since AI reliability is capped by how much ambiguity still exists in the business it's operating inside.

About Jason HartJason Hart is a marketing data analytics and AI expert, multi-time SaaS founder, and analytics engineer, and founder of Domain Methods, focused on revenue data pipeline architecture and AI reliability engineering.

Key Timestamps / Chapter Markers[0:04] Introduction and episode framing[4:18] What AI is actually doing when it generates an output[8:42] What most people get wrong about AI reliability[12:12] What happens when probabilistic AI meets deterministic systems[16:37] Where AI systems break moving from demos into production[20:06] How messy data infrastructure undermines AI performance[28:26] Where AI hype most clearly diverges from reality[33:18] Integrating AI into decentralized systems without losing trust

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

For decades, scale meant one thing: more people, more layers, more structure. But that assumption is cracking. Small teams, sometimes just one or two people, are building products, moving capital, and competing with organizations that used to require hundreds of employees — not by working harder, but by being structured differently.

Danny Carpio is an organizational architect and systems builder who has helped design and operate companies and networks that raised over $25 million in capital and reached nine-figure valuations. He is the author of The Unifirm, a framework built around a disruptive idea: the new unit of scale isn't the company, it's the individual.

In this episode, Collins Victory Odabi sits down with Danny to explore why the traditional firm model is breaking down, what "self-scaling" actually looks like in daily practice, and whether Web3 is genuinely building new organizational forms or just rebranding old ones.

In This Episode, We Discuss

• The firm isn't broken — it's no longer the optimal organizing mechanism. Danny draws on Coasian transaction-cost theory to argue firms exist purely because certain functions were cheaper to do inside one; as AI collapses those costs, the structures built around scarcity are being exposed as inefficient, not evil.

• Self-scaling starts with radical, unflattering self-assessment, not more tools. Danny insists you must "datafy yourself" — honestly mapping what you're genuinely good at, what you only think you're good at, and what the market will actually pay for, before building any system around your time.

• Return on time, not hours worked, is the core metric of a self-scaled operator. Danny treats his time like a portfolio asset, filling it only with activities that compound, and structures relationships around generating outsized value for the people he works with rather than transactional exchange.

• AI doesn't create good outcomes — it accelerates whoever is already good at something. Danny warns that AI can make you feel skilled at things you aren't, and argues the real competitive edge is shifting toward design, judgment, and strategic patience rather than speed of execution alone.

• Web3's real organizational future isn't decentralization for its own sake — it's composability. Danny argues DAOs often failed by discarding what firms did well instead of rebuilding it, and that the next wave of crypto-native organizations will succeed by treating individuals as interoperable "nodes" the market can freely plug into.

About Danny CarpioDanny Carpio is an organizational architect, systems builder, and author of The Unifirm, with experience across Ready Player DAO, Hydra Ventures, and his current AI-native startup, Very.

Key Timestamps / Chapter Markers[0:04] Introduction and episode framing[2:16] From law school to Chainlink, DAOs, and building Hydra Ventures[7:06] What's actually broken about the traditional firm model[10:57] The economic shifts that turned firms into a constraint[14:48] What self-scaling actually looks like day to day[22:05] How working inside DAOs shaped his thinking on organizational design[26:46] The biggest misunderstanding about AI's role in how work gets done[36:15] Is Web3 building new organizations, or just rebranding old ones?

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

If AI agents can act, can they also earn? If they can earn, can they own? And if they can own, can they become independent economic entities? That question sits at the intersection of two disruptions colliding at once, and the gap between imagining that future and actually building it introduces real constraints around trust, security, governance, and value creation.

Andrew Melnychuk is an early Bitcoin investor, angel investor, and founder building an AI agent system that integrates crypto payments, exploring what he calls agentic capital markets — a model where agents don't just execute tasks but operate as participants in economic networks.

In this episode, Collins Victory Odabi sits down with Andrew to explore what his working product actually does today, why he believes crypto is market infrastructure rather than just a payments layer, how agents could eventually raise their own capital, and the failure points he's most worried about.

In This Episode, We Discuss

• Andrew has a working, revenue-generating product, not a prototype. His "hyperagent" builds other agents — one example pulls SEO data from multiple platforms and produces a report that used to cost $1,500 and take two weeks, now completed in minutes for a few dollars.

• Crypto isn't a payment layer bolted onto agents — it's the market infrastructure that lets them exist as economic actors. Andrew argues cryptocurrency solves the custody and coordination problems traditional finance built clearinghouses to handle, and that agents will use this infrastructure to transact, hold equity, and pay dividends without ever needing a bank account.

• Agents could eventually "IPO" themselves through instruments like Bitcoin's Taproot assets. Rather than a traditional IPO costing millions, Andrew describes an agent creating tokenized equity, exchanging it for capital, and paying dividends from work it performs — with full financial reporting handled automatically.

• Andrew doesn't believe AI agents will replace human workers, and calls the "replacement" narrative genuinely irresponsible. He insists agents require constant human supervision, describing them as powerful tools for retrieving, ranking, and recommending — not independent intelligence, despite what's commercially convenient to claim.

• His biggest fear is a fully autonomous agent with its own compute, energy, and crypto wallet acting outside anyone's control. Andrew describes a scenario where an unsupervised agent builds real-world infrastructure for harmful ends — arguing the real safeguard isn't shutting the technology down, but building in insurance, signed accountability, and mechanisms to verify who's actually responsible for an agent's actions.

About Andrew MelnychukAndrew Melnychuk is an early Bitcoin investor, angel investor, and founder building AI agent systems integrated with crypto payments, focused on the emergence of agentic capital markets.

Key Timestamps / Chapter Markers[0:05] Introduction and episode framing[2:29] From fine arts and Star Trek to Bitcoin and AI agents[8:34] What Andrew's hyperagent product actually does today[13:41] Why crypto is market infrastructure, not just a payment layer[18:24] How an agent could "IPO" itself and raise capital[26:40] The biggest technical and economic failure points so far[35:54] Balancing agent autonomy against governance and control

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

One of blockchain's biggest challenges has never been technology — it has been relevance. For years the industry has searched for its killer application while countless projects promised to revolutionize everything, yet many never made it past speculation because they started with blockchain rather than an actual business problem.

Dan Beech is an entrepreneur, investor, and founder of Beachworks and Archively. Across more than two decades, he has built and led companies spanning real estate, wealth management, healthcare, lending, consumer products, and blockchain, scaling organizations generating tens of millions in annual revenue.

In this episode, Collins Victory Odabi sits down with Dan to explore why blockchain succeeds only when it solves meaningful business problems, why document storage may be one of the technology's most overlooked applications, and what founders can learn from building in highly regulated industries.

In This Episode, We Discuss

• Archively was born from a gap Dan saw running a wealth management firm, not from chasing blockchain hype. He noticed the chaotic space between when a family leader dies or is incapacitated and when the family can move on peacefully — a gap that exists even with good estate planning, and one built on document storage infrastructure that's already 20 years old.

• Dan explains blockchain to skeptical executives by starting with something familiar: cloud storage. He describes it as your document mirrored across 10,000 servers worldwide, constantly cross-checking each other for changes — meaning even if half those servers go down, your documents stay accessible.

• The first rule for any founder is solving a real problem, not searching for one to fit the technology. Dan's clearest advice: talk to potential users first, understand their actual pain points, and only then map where your technology can solve them — never build a solution in search of a problem.

• Archively splits documents into emergency access and long-term family access, and blockchain solves both differently. Emergency documents (a medical directive, a passport copy) need to survive outages like the AWS or Outlook failures Dan cites; long-term documents (a will, final wishes) need infrastructure built to last decades, which no single company can guarantee for itself.

• Archively runs on a real-time mirror to the blockchain, not a backup. The site pings itself every five minutes; if it goes down, the blockchain-backed version automatically activates — slower, but functional exactly when a family might need a document most urgently.

About Dan BeechDan Beech is an entrepreneur, investor, and founder of Beachworks and Archively, with more than two decades of experience building companies across real estate, wealth management, healthcare, lending, and blockchain.🌐 archively.com

Key Timestamps / Chapter Markers[0:05] Introduction and episode framing[3:39] The operational gap that led to building Archively[7:17] How Dan explains blockchain to non-technical business leaders[10:34] Operational disciplines founders consistently underestimate[13:36] What Archively can do that traditional cloud storage can't[20:23] The first step before investing time or money into a new venture

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

For most of modern history, financial education has followed a predictable script: earn, save, invest in well-understood assets, repeat. But today's landscape includes crypto, tokenized assets, and decentralized tools alongside traditional ones — lower barriers to entry, but far higher complexity. That raises an urgent question: how should we teach the next generation about money?

Chandler Walker is a father, entrepreneur, and creator of Culture of Care, a fast-growing newsletter helping parents raise financially fluent kids and build generational wealth within the family. After scaling multiple seven-figure companies and coaching thousands of founders, Chandler shifted focus to practical, early financial exposure for children over abstract theory.

In this episode, Collins Victory Odabi sits down with Chandler to explore why most family wealth disappears within two or three generations, how he built a "family office" system for his own household without an army of lawyers, and how his daughter learned to invest by age six.

In This Episode, We Discuss

• Most family wealth disappears within 2-3 generations because trusts aren't operating systems. Chandler argues even well-structured trusts fail without a repeatable framework — so he built his own family office around growth, protection, and speculation, designed so his daughter could execute it immediately if something happened to him.

• Parents consistently underestimate what young children can understand. Chandler's daughter began investing in familiar companies like Roblox at six, learning to deploy money without checking it obsessively — and learned about loans and interest by borrowing against points earned for good behaviour.

• Every asset should fall into one of three buckets: growth, protection, or speculation. Chandler applies this across traditional markets, crypto, and Web3 — a stablecoin earning yield sits in protection, Bitcoin sits in long-term growth, and a token like Uniswap, chosen by his daughter for its unicorn logo, sits in speculation she expects could go to zero.

• Younger generations have easier market access — and lose money faster because of it. Chandler points to apps like Robinhood and Kalshi giving retail traders speed and access once reserved for institutions, which he expects will eventually push more people toward structured systems after losing money to unstructured speculation.

• The single most important financial skill for the next 20 years is a "systems mindset." Chandler's litmus test: if you can't explain in one sentence why you own an asset and where it leads, you don't understand it well enough to hold it — a discipline he argues prevents the FOMO-driven decisions that get people into trouble in Web3.

About Chandler WalkerChandler Walker is an entrepreneur and creator of Culture of Care, a newsletter helping parents raise financially literate children and build generational wealth through structured family systems.📱 Instagram: @Chandler_SAF

Key Timestamps / Chapter Markers[0:05] Introduction and episode framing[2:42] From med school to a gym to buying Ethereum at $15[4:58] Why Chandler built his own family office system[7:47] Common mistakes parents make introducing kids to money[10:36] What his daughter's first stock purchase revealed about kids and ownership[12:24] Are the principles of wealth building changing, or just the tools?[20:15] Introducing children to digital assets alongside traditional ones[28:41] The most important financial skill for the next 20 years

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

There's a dangerous assumption founders make: if money is coming in, the business is working. Revenue becomes validation, growth becomes proof, activity gets mistaken for stability. But revenue is not control — you can be doing seven figures and still not know where your money is going, growing while structurally unstable, until volatility exposes it fast.

Daniel Bamper is a chemist turned CFO and founder of Black Belt Ledgers, where he helps businesses install financial systems that turn chaotic revenue into structured profits. His work focuses on identifying hidden leaks, fixing broken financial processes, and helping operators move from reactive decisions to controlled systems.

In this episode, Collins Victory Odabi sits down with Daniel to explore why revenue and profitability aren't the same thing, the biggest financial blind spots operators hit while scaling, and what a real financial control system actually looks like.

In This Episode, We Discuss

• Businesses need faster financial feedback as they scale, not just more of it. Daniel compares early-stage business to walking through a city with a map, but fast growth is like driving a sports car through it — you need real-time GPS, not a paper map, or you lose your bearings entirely.

• Profit leaks hide in materials, labour, and subcontractor spend. Daniel describes trade businesses losing money to inefficient job-site ordering and uncoordinated crews, and stresses that a simple efficiency ratio — hours billed versus hours paid — often reveals technicians operating at 40% when 50–60% is achievable.

• The right financial system scales with revenue, not with ambition. Daniel lays out concrete thresholds: a spreadsheet under $100K, an outside bookkeeper with monthly reporting past that, weekly reporting past $500K, plus a "profit model" — tracking profitability by project or by program depending on the business type.

• Unpredictable income is usually a symptom of an inconsistent sales system, not bad luck. Daniel argues real financial stability requires action targets, not just revenue targets — tracking how consistently leads move from unknown to prospect to closed client, since money is often "a consequence" of a broken pipeline.

• Every role in a business should own a measurable number. Citing the book Traction, Daniel describes how assigning ownership of specific performance metrics — down to something as granular as answering the phone within two rings — is what makes volatility visible before it becomes a crisis.

About Daniel BamperDaniel Bamper is a chemist turned CFO and founder of Black Belt Ledgers, helping seven-figure trade and contracting businesses build financial systems that scale into eight figures.

Key Timestamps / Chapter Markers[0:04] Introduction and episode framing[2:01] From organic chemist to martial arts school owner to CFO[5:58] Why significant revenue doesn't guarantee cash flow or profitability[8:53] The biggest financial blind spots operators hit while scaling[11:46] What a real, in-control financial system looks like[14:39] Building financial stability when income is unpredictable[18:58] Identifying and eliminating financial leaks before they compound[22:54] Where to start moving from reactive decisions to structured control

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

There's a pattern repeating across emerging technology: new tools appear, adoption accelerates, capital flows in — then systems break, not because the technology fails, but because the foundations weren't built to handle scale, risk, or complexity. AI is deployed without governance, blockchain systems run unclear security models, and leaders make high-stakes technical decisions without fully understanding the systems they're responsible for.

Kelvin Carlson is a tech CXO partner, five-time CTO and CISO, and executive coach who helps companies scale technology and leadership systems simultaneously. His work sits at the intersection of cybersecurity, AI governance, infrastructure design, and executive decision-making.

In this episode, Collins Victory Odabi sits down with Kelvin to explore how sitting across CEO, CTO, and CISO roles changes how problems get diagnosed, why security failures remain the most common "fire" companies call him in for, and what real AI governance looks like beyond policy documents.

In This Episode, We Discuss

• Companies rarely fail from one bad decision — they fail from neglecting balance too long. Kelvin describes how sales-, product-, or engineering-driven organisations can each tip too far for extended periods, and that the real fix is leadership open to honest pushback from every function.

• Security has been the most common "fire" of the past five years. Kelvin points to younger companies deprioritising security until a vulnerability gets exploited or data gets exfiltrated, and stresses that basic hygiene — like GitHub's Dependabot alerts — often sits unused.

• Carelessness, not malice, is the most dangerous AI integration mistake. Kelvin identifies two forms: submitting sensitive data (like HIPAA information) to non-compliant LLMs, and running coding agents with unrestricted permissions — citing a case where an AI agent deleted a company's database and backups in nine seconds because no one was watching.

• Real AI governance means codified guardrails, not just policy documents. Kelvin describes filtering prompts and outputs at the code level, guarding against prompt injection, and logging every prompt for audit — treating LLM oversight as infrastructure, not a compliance checkbox.

• A company's real security footprint extends far beyond its own code. Kelvin argues resilience must account for every third-party tool, API, and SaaS integration a company relies on, noting employees and contractors are often the most fragile point of entry.

About Kelvin CarlsonKelvin Carlson is a tech CXO partner at TechCXO, a five-time CTO and CISO, and an executive coach specialising in cybersecurity, AI governance, infrastructure design, and executive decision-making.

🔗 LinkedIn: linkedin.com/in/KevinCarlson📰 Newsletter: The Executive Shift (on LinkedIn)🌐 techcxo.com

Key Timestamps / Chapter Markers[0:05] Introduction and episode framing[2:16] Kelvin's path from a 1970s HP 3000 to five CTO roles[4:46] How sitting across CEO, CTO, and CISO roles changes diagnosis[8:04] The most common "fire" companies call Kelvin in to fix[10:31] The most dangerous mistake companies make integrating AI[14:37] What real AI governance looks like beyond policy documents[18:31] Rethinking cyber resilience for decentralised, Web3 systems[22:11] Why companies are turning to fractional CTOs and CISOs

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

We live in an era where technology evolves faster than our ability to understand its long-term consequences. AI is transforming decision-making, Bitcoin continues to challenge conventional ideas about money, and governments spend billions on ambitious programs with wildly different outcomes. Beneath these separate stories lies a common thread: systems, the incentives they create, and the assumptions they quietly reward or punish.

Alberto Artasanchez is a data scientist, entrepreneur, solutions architect, and author whose career spans more than three decades across artificial intelligence, cloud computing, finance, investing, and real estate. Whether writing about Bitcoin, AI, or why billion-dollar public projects fail, he approaches every subject from the same angle: understand the system before believing the story.

In this episode, Collins Victory Odabi sits down with Alberto to explore what AI can and cannot do for investors, why Bitcoin demands a different way of thinking, and what repeated institutional failures teach us about incentives, execution, and the limits of technology.

In This Episode, We Discuss

• Markets, not boredom, usually decided when Alberto changed direction. Reflecting on exiting a real estate portfolio, Alberto describes how a mortgage rate reset from 3% to 7%, alongside new rent-control laws, made the decision for him — reading what a system is telling you matters more than personal conviction.

• AI's real value in investing is finding edges, not answering "should I buy this stock." Alberto uses AI to generate trading algorithms that identify and exploit market inefficiencies, arguing the technology's usefulness depends entirely on the sophistication of the questions and prompts behind it — and that today's use case will look outdated within months.

• Bitcoin's scarcity is enforced by energy, not belief. Unlike currencies or gold, Alberto argues Bitcoin's 21 million cap is validated by the real, unfakeable energy cost of running its network — making it near-impossible to replicate, and positioning it, in his view, as the eventual medium of exchange between AI agents themselves.

• California's high-speed rail failed because its incentives were designed to fail. Alberto breaks down how the project's funding structure — a promise of no subsidies and a capped $50 fare — guaranteed private investors had no reason to fund it, since there was no return to chase. The failure wasn't incompetence; it was incentives working exactly as structured.

• Alberto's two-step formula for achieving anything: find people who've done it, and ignore everyone who hasn't. His closing advice is to seek out and follow those with direct, demonstrated experience in what you're trying to achieve, while deliberately disregarding well-meaning advice from people who have never actually done it themselves.

About Alberto ArtasanchezAlberto Artasanchez is a data scientist, entrepreneur, solutions architect, and author with more than three decades of experience across artificial intelligence, cloud computing, finance, investing, and real estate.

Key Timestamps / Chapter Markers[0:05] Introduction and episode framing[2:10] The one lesson from jumping across industries and business cycles[7:27] Where AI genuinely improves investment decisions — and where it doesn't[11:08] Why Bitcoin is a decades-long game, not a day trade[15:55] What a $42 billion broadband failure teaches about incentives[21:07] Do decentralized systems really solve structural problems?[24:00] Final thoughts: the two-step formula for achieving anything

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.

Artificial intelligence has moved quickly from novelty to necessity, yet many companies remain stuck in an awkward middle ground. Employees experiment with ChatGPT, teams test isolated workflows, and leaders invest in subscriptions — but very little becomes embedded in how the business actually functions. The problem isn't access to AI anymore. It's implementation.

Marnie Wills is an entrepreneur, AI strategist, and trainer who helps non-technical founders and business leaders move beyond scattered experimentation towards practical AI systems that genuinely improve how they work. Having built and exited businesses herself before embracing AI, she understands both the hesitation founders feel and the operational work required to make AI useful rather than overwhelming.

In this episode, Collins Victory Odabi sits down with Marnie to explore why AI adoption fails inside many organisations, how non-technical leaders can implement it successfully, and why the future belongs to those building the best AI workflows rather than using the most tools.

In This Episode, We Discuss

• The biggest AI adoption myth is that it requires technical skill. Marnie argues the fastest win is a mindset shift — treating a large language model as a thinking partner that increases human capability, not a technical system requiring expertise, whether that's Copilot, Gemini, or ChatGPT.
• Most businesses try to force AI into fragmented old systems. Marnie calls this fitting "a square peg into a round hole" — running one AI workflow alongside dozens of disconnected tools instead of building agents with shared context across the business.
• Automation isn't the starting point — clarity is. Before automating anything, Marnie insists founders map their actual step-by-step processes, since an AI can't execute a workflow that only exists in someone's head. She works through lead gen, client delivery, and admin, finding real bottlenecks before building any agentic workflow.
• The "human-in-the-loop" concept is widely misunderstood, fuelling an AI slop problem. Marnie proposes a 20-60-20 model: 20% domain expertise going in, 60% AI doing the heavy lifting, 20% critically challenging the output. She cites that 52% of websites launched in June 2025 were AI-generated or hybrid, warning that without real human expertise embedded, future AI has nothing authentic left to learn from.
• Underused built-in AI tools like Microsoft's Work IQ and Google's personal intelligence are hiding in plain sight. Marnie describes using Copilot inside email and Teams, and Gemini to identify a car tyre from a photo and find a nearby garage — the real opportunity isn't new tools, but using the AI already embedded in daily tools.

About Marnie WillsMarnie Wills is an entrepreneur, AI strategist, and trainer who helps non-technical founders and business leaders build practical AI systems that genuinely improve how they work, drawing on her own experience building and exiting businesses.

📱 LinkedIn and Instagram: @businesswithaistrategist

Key Timestamps / Chapter Markers[0:04] Introduction and episode framing[2:07] Why implementation, not technical skill, is the real AI barrier[5:46] Fixing fragmented systems before automating anything[10:25] Deciding what to automate without unnecessary complexity[14:23] The human-in-the-loop concept and the AI slop problem[19:45] Using built-in AI like Copilot and Gemini to get started[23:20] Final thoughts and where to connect with Marnie

Connect with CollinsLinkedIn: linkedin.com/in/collins-odabi-266620380Guest enquiries: podmatch.com/member/web3unfiltered

Until next time, stay sharp, stay curious, and keep building.