About
I help financial institutions turn business challenges into measurable transformation…
Articles by Steve
Activity
12K followers
-
Steve Tetrault shared this70% in five years. That was the transformation challenge embedded in a recent RUN operations RFP. The wrong response would have been to propose another labor-arbitrage program, move more tickets to a lower-cost delivery center, or attach an AI copilot to every queue. You cannot labor-arbitrage your way to structural transformation. You have to change the work itself. That means: • Eliminating unnecessary operational demand • Preventing incidents before they become tickets • Standardizing platforms and delivery paths • Bringing spec-driven development into the AI SDLC • Turning runbooks and tribal knowledge into reusable intelligence • Governing AI, data, security, identity, and cost as one operating system • Connecting every technical signal to a business outcome My latest article lays out a Crawl. Walk. RUN model: CRAWL — See and stabilize WALK — Standardize and automate RUN — Predict, prevent, and self-optimize The future of RUN is not ticket management. It is a continuously learning business-performance system. Business as usual is no longer business as usual. #RunOps #AIOps #SRE #PlatformEngineering #FinOps #AISDLC #SpecDrivenDevelopment #DigitalTransformation
-
Steve Tetrault shared thisThe laundering network already knows your org chart. KYC sees the customer. Sanctions sees the name. Fraud sees the scam. Cyber sees the device. Payments sees the movement. AML sees the pattern. Investigations see the case. The criminal network sees one system. That fragmentation is not just an operating inconvenience. It is part of the attack surface. Part III of my CAMS-inspired series argues that the defensive unit must move: • The alert • The customer cluster • The beneficiary network • The wallet cluster • The mule structure • The typology • The investigation hypothesis AI can help with entity resolution, network analysis, chronology, source reconciliation, missing evidence, and draft narratives. It should not determine guilt, file autonomously, exit customers, adjudicate sanctions, or disclose protected information. A model inference must never become a fact merely because it appears inside a persuasive case narrative. The operating loop is: • Detect • Connect • Investigate • Report • Share • Disrupt • Learn The SAR is an important intelligence product. It is not the final outcome. The outcome is better intelligence, faster and more defensible action, stronger controls, and less successful illicit finance. Read the article below. #AML #CAMS #FinancialCrime #TransactionMonitoring #KYC #Sanctions #AgenticAI #FinancialServicesThe Laundering Network Already Knows Your Org Chart.The Laundering Network Already Knows Your Org Chart.
-
Steve Tetrault shared thisBefore you go all in on AI, know the business you’re betting on. ♠️ I’ve already invited you to our All In at ITC 2026 poker night. Here’s what else I’m bringing to the table: my newly earned LOMA Associate, Life Management Institute (ALMI) designation. Because knowing the technology gets you into the conversation. Knowing the industry gives you something worth saying.. Here’s what else I’m bringing to the table: my newly earned LOMA Associate, Life Management Institute (ALMI) designation. Yes, another acronym. No, that’s not the point. The point is deeper understanding of underwriting, product economics, distribution, policy administration and claims—not just the technology we put around them. Knowing how to build AI isn’t the same as knowing where it belongs. I keep investing in industry expertise because my clients deserve both: technical depth and a genuine understanding of their business. Better questions. Better judgment. Better outcomes. I want faster underwriting without weaker risk decisions. Automation that gives advisors more time for people. Claims experiences that make a beneficiary’s hardest day easier—not turn it into a chatbot scavenger hunt. That’s the conversation I’m bringing to ITC: less “look what the model can do,” more “show me what changed for the business and the customer.” And yes, we can have it over a few hands of poker. Join IBM and Microsoft at All In at ITC 2026 on September 30, at 7 p.m. PDT, in Las Vegas. 🃏https://lnkd.in/gwG54QyC My poker face is questionable. My commitment to learning isn’t. #ITC2026 #ALMI #LOMA #LifeInsurance #AITransformation
-
Steve Tetrault shared thisYour customer passed KYC. The network didn’t. A clean onboarding file can hide: • A dirty ownership graph • A new controller • Synthetic identity • Sanctioned exposure • A shared device • A wallet outside the customer master • Behavior that changed yesterday • A software agent acting under unclear authority Point-in-time KYC assumes the customer remains substantially the same until the next scheduled review. That assumption is getting harder to defend as ownership, channels, counterparties, wallets, devices, and delegated agents change continuously. Part II of my CAMS-inspired series argues for a living customer-intelligence layer: • Identity • Ownership and control • Screening • Behavior • Relationships • Case history • Source-linked evidence AI can retrieve, reconcile, summarize, recommend, and prepare. It should not autonomously change a risk rating, approve EDD, exit a customer, adjudicate sanctions, file a SAR, or share protected information. The result should not be one mysterious score. It should show what is known, what changed, what conflicts, what is missing, and who approved the conclusion. The customer-risk rating is not a fact. It is a time-bound conclusion supported by evidence. The future of KYC is not more frequent paperwork. It is a continuously governed understanding of the customer and the network around them. Read the article below. #KYC #AML #CAMS #FinancialCrime #DigitalIdentity #Sanctions #AgenticAI #FinancialServices
-
Steve Tetrault shared thisCAMS certified. Same loud jacket. Sharper perspective. 🐆 I’ve earned my Certified Anti-Money Laundering Specialist (CAMS) designation from ACAMS. CAMS_Certificate_Steve Tetrault_ER-005247235.pdf The journey inspired my three-part AML series and one question: Are we reducing financial crime risk—or just getting better at checking boxes? As a technology leader, I want to bring that lens to AI: better context, stronger decisions, and measurable impact—not just more automation. The certificate is a milestone. Applying the knowledge is the point. #CAMS #ACAMS #AntiMoneyLaundering #FinancialCrime #AI
-
Steve Tetrault shared thisThere’s enough bluffing in tech. Let’s do it properly. 🃏 I’m joining the IBM and Microsoft teams for some "All In" action at ITC 2026—bringing insurance and technology leaders together for an evening of poker, conversation, and a little friendly competition. The only deck you will see has 52 cards. Let’s get past the polished talking points. What’s actually working in insurance and AI? What’s still hype? And who’s been confidently playing a terrible hand? That’s the conversation I’m looking forward to—with some laughs, great food and drinks and unbelievable prizes, new connections, and a chance to find out whether your poker face is as convincing as your business case. Bring your opinions. Bring your competitive streak. 📍 Las Vegas 📅 September 30, 2026 | 7 PM PDT Seats are limited. Registration and confirmation are required. 🔗 Reserve your seat https://lnkd.in/gwG54QyC Who’s in? #ITC2026 #InsureTechConnect #InsuranceLeaders #IBM #Microsoft
-
Steve Tetrault shared thisYour AML program passed the audit. The criminal network passed through it. That is the uncomfortable gap between compliance activity and program effectiveness. An institution can prove that it: • Approved the risk assessment • Reviewed the alerts • Closed the cases • Filed the SARs • Completed the training • Tested the program Then struggle to answer: Which illicit-finance risks actually declined? Which controls created useful intelligence? Which scenarios should be retired? Which customer segments consume capacity without corresponding value? Which investigations changed the next decision? As I begin the CAMS journey, I am exploring where the established AML control model meets AI, digital assets, new regulatory expectations, and networked financial crime. Part I argues for a harder operating model: Risk Assessment → Customer Understanding → Monitoring → Investigation → Reporting → Information Sharing → Learning AI can help retrieve, reconcile, summarize, test, and recommend. It cannot own guilt, SAR filing, customer exit, sanctions adjudication, or protected information sharing. The next AML program should not produce more paperwork faster. It should produce better intelligence, more defensible decisions, and measurable control improvement. Read the article below. #AML #CAMS #FinancialCrime #AgenticAI #Compliance #RiskManagement #KYC #FinancialServicesYour AML Program Passed the Audit. The Criminal Network Passed Through It.Your AML Program Passed the Audit. The Criminal Network Passed Through It.
-
Steve Tetrault shared thisFraud will become agentic before your bank does. Criminals do not need: • A model-risk committee • A change window • A complete data-governance program • Approval to test another variant • Permission to scale INTERPOL now warns that agentic AI can plan and execute complete fraud campaigns. The real shift is not simply better deepfakes. It is coordination: Reconnaissance → Identity construction → Social engineering → Account opening → Transaction execution → Mule routing → Laundering → Adaptation Most controls still see the individual events. The attacker sees the campaign. In the final article of Fraud at Machine Speed, I outline the defensive response: 1. Build campaign graphs—not bigger alert queues. 2. Treat identity as a composite system. 3. Connect real-time and retrospective intelligence across entities and time. 4. Give investigators campaigns, evidence, and rationale. 5. Bound the authority of defensive agents. 6. Treat lawful information sharing as part of the control system. 7. Operate a closed loop: Detect → Discover → Interdict → Investigate → Learn. The outcome is not more alerts. It is less successful fraud. Read Part III below. Part I: Your Fraud Alert Is a Postmortem https://lnkd.in/gF-pSWRr Part II: Your Fraud Model Is a Very Expensive Inbox #FraudAtMachineSpeed #AgenticAI #FraudPrevention #FinancialCrime #AML #FinancialServices #MicrosoftAzure #CyberSecurity
-
Steve Tetrault shared thisYesterday was my birthday, and of all the thoughtful wishes I received from family and friends, the sweetest surprise came from GARP: I passed Part I of the Financial Risk Manager (FRM) exam. First, thank you to everyone who called, texted, posted, or reached out. I genuinely appreciated every birthday message. This result made the day even more memorable. Over the last 20 years, I have sat for approximately 50 professional certification exams across technology, cloud, architecture, AI, financial services, and insurance. FRM Part I was exponentially the hardest. The breadth of the curriculum, quantitative rigor, time pressure, and need to connect concepts across risk foundations, financial markets, products, valuation, and risk models made this unlike any exam I had taken before. But the difficulty is only part of why this milestone matters. I have built my career at the intersection of business, technology, and transformation. To deliver exceptional outcomes for banks, capital-markets firms, and insurers, it is no longer enough to understand cloud platforms, data, AI, and architecture. We must also understand the environment in which those technologies operate: Risk. Capital. Markets. Financial products. Controls. Regulation. Operational resilience. Customer and policyholder trust. The FRM journey is a deliberate investment in building that depth. As AI becomes embedded in underwriting, claims, fraud detection, financial crime, risk management, servicing, and core operations, technology leaders must understand both what the technology can do and the consequences of the decisions it enables. We need to connect: Innovation to risk appetite Architecture to enterprise controls Automation to accountability AI decisions to transparency and governance Technology investment to measurable customer and business outcomes That deeper domain expertise allows me to ask better questions, engage more meaningfully with business and risk leaders, design more resilient solutions, and help clients pursue transformation in ways that are innovative, defensible, and worthy of trust. My commitment to continuous learning has never been about collecting credentials. It is about continually expanding the value I can bring to my clients, colleagues, and teams. Customer excellence begins with doing the work to understand the customer’s world—not only their technology, but their business, products, risks, obligations, and ambitions. Part I represents meaningful progress, but it is not the finish line. Part II is next. Grateful for the birthday wishes. Proud of this milestone. And more committed than ever to learning, growing, and raising the standard of what I bring to every client engagement. #FRM #GARP #FinancialRiskManagement #FinancialServices #Insurance #RiskManagement #AITransformation #ContinuousLearning #CustomerExcellence #Leadership #IBM https://lnkd.in/g7EwVxnE
Steve Tetrault replied to a comment
1mo
Kunaal Kapoor Thank you for the leadership and inspiration. In full transparency the one that is taking a lot of time is the FRM® | Financial Risk Manager Certification | GARP. I have written Part I and Part II is scheduled for next spring.
-
Steve Tetrault liked thisSteve Tetrault liked thisI've just completed the Celonis Foundations Training Track - a lot more to come! #ProcessMining #processintelligence #CelonisAcademy #IBMConsultingCelonis Foundations was issued by Celonis to Shouryasimha Addepalli.Celonis Foundations was issued by Celonis to Shouryasimha Addepalli.
-
Steve Tetrault liked thisSteve Tetrault liked thisGreat experience at the US Open today with clients, partners, and some amazing IBMers. Excellent tennis 🎾 and great to see Osaka and Fritz wins. Was awesome to use IBM powered fan experience throughout but also to see behind the scenes technology that touches ~15M at home and ~1M stadium fans. Congrats to IBM sports team that ran the show with industry leading #ai, #hybridcloud, real time #data, and #security. That is what production AI looks like when it's real: boring, reliable, and invisible to the customer.
Experience & Education
Licenses & Certifications
Organizations
-
Azure Advisory Board
Advisor
- Present -
Professional Association Of SQL Server
Member
- Present
Recommendations received
2 people have recommended Steve
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content