Bradley B.

We're facing more and more questions about the use of synthetic data in market research. I'd recommend this blog post, which really resonated with me. So much of the conversation is "how closely can this replicate human data?" But there are so many other considerations researchers should be thinking about. https://lnkd.in/eSQfSTPj

31

3 Comments

Consider a borderline lending case. A score shifts from 0.49 to 0.51. The outcome flips from decline to approve. The real question is not accuracy.... It is whether the institution can demonstrate: -which data fields were used and which were excluded -which policy constraints were enforced at runtime - which model version was active - why the threshold was set where it was - what the counterfactual outcome would have been - what evidence was preserved for Model Risk Management review Prediction is statistical;Decision-making is governance. When those layers are fused into opaque workflows, risk accumulates quietly and surfaces later in audit findings, remediation cycles, or litigation. Institutions that separate prediction, policy enforcement, and evidence preservation move faster with less downstream friction.

7

4 Comments

🔹 Data Governance: The Work That Moves Revenue 🔹 People often say data stewardship isn’t flashy. I couldn’t disagree more. Every taxonomy I define, every usage rule I enforce, and every governance process I refine creates downstream trust in data. That trust is the foundation of analytics, compliance, and innovation. Without it, business decisions are guesses—not strategy. 💡 When governance is done right, it’s not about control—it’s about enabling growth. 💡 When data is trusted, it directly fuels revenue, efficiency, and competitive edge. 💡 When stewardship is strategic, businesses thrive. I take pride in being the “revenue mover behind the scenes,” because I know the work I do makes the difference between data chaos and data clarity. 👉 How does your organization demonstrate the business value of governance work to stakeholders who don’t see the daily grind? #DataGovernance #DataStewardship #DataQuality #AnalyticsLeadership #DigitalTransformation #DAMA

2

1 Comment

If your data/analytics team isn't leading the charge on AI, you're doing your organization a disservice. Tris J Burns 🔥 wrote about this in his newsletter and I wholeheartedly agree. Let’s be clear... it’s not because AI is going to replace data analysts (or anyone else). So why? Two really good reasons: 1. Someone, somewhere in your organization is using it, and they need your help. Data teams know the data, know the limitations and the nuances, how to be rigorous, and how rigorously we need to validate data before drawing conclusions . 2. Data teams know where the bodies are buried. All those weak data pipelines, poor metadata and workarounds you've been complaining about for years impact AI. No-one knows how to fix them better than you. I have many reservations about AI and the conclusions it draws, but used right, it's a fantastic tool. I use it, I've weaseled my way into several AI projects (mostly to annoy Peter G. McDermott), and regularly share my perspective on what we need to do to our data to get it working right. As a data person, you should be doing the same, otherwise your AI strategy will be built on weak foundations. What is your team doing to leverage AI or support AI strategy?

84

17 Comments

The biggest blocker to analytics success isn’t bad data. It’s pride. The data is clean. The model works. The story is clear. The real challenge comes next: change. People take pride in their work. They’re the expert. So when data shows a better way, it can feel like a threat instead of a tool. But analysis doesn’t assign blame. It shows a path forward. It’s not about failure. It’s about progress. It’s not “you got it wrong.” It’s “here’s the next step.” Data doesn’t diminish expertise. It strengthens it—if you let it.

23

We are leaving retrieval performance on the table by treating documents as flat sequences of text. When you split a #financial report or a technical manual every 500 words, you inevitably cut through the middle of an idea. you might separate the problem from the solution just because the word count reached a limit. This leads to retrieval failures where the model misses the context entirely. A recent approach regarding Hierarchical Text Segmentation proposes a smarter, bottom-up #architecture that solves this without inflating costs. Instead of blind cutting, the model functions like a human reader. First, it identifies natural semantic breaks where the topic actually changes (#segmentation). Second, it groups these related segments into broader themes (#clustering). The real value for business applications is in the "Dual-Vector" retrieval strategy. The system indexes both the specific segment and the broader cluster. If a user asks a specific question like "Who signed the audit?", the system matches the specific segment #vector. If they ask "What is the overall risk profile?", it matches the broad cluster vector. You get the best of both worlds: granular precision and high-level context. #RAG #chunking #GenAI #AIEngineer #Banking #ML

5

Reporting should be the fastest part of treasury. 𝘕𝘰𝘵 𝘵𝘩𝘦 𝘴𝘭𝘰𝘸𝘦𝘴𝘵. → But teams are still pulling from disconnected systems → Still reconciling manually → Still rebuilding logic every month just to get a baseline report out This isn’t a tooling problem. It’s an infrastructure problem. If you don’t have a consistent source of truth—and a way to explore it live—you’ll always be chasing numbers instead of using them. What’s still slowing down your reporting today? #Treasury #CorporateTreasury #Treasury4 #FinanceTransformation #DataInfrastructure #FinancialReporting #CFO #FinTech #TreasuryTech

20

1 Comment

We’ve spent 20 years trying to perfect pattern matching, but we’re still hitting the "Intelligence Wall." It’s a wake-up call for anyone building agentic workflows today. They argue that true autonomy doesn't come from data—it comes from a hierarchy of abstraction that ends in "Tricks." The Research Reality Check: The CBR vs. EBR Gap: Most RAG systems today are just Case-Based Reasoning (CBR). They retrieve and adapt. But the research points toward Experience-Based Reasoning (EBR)—a meta-inference layer that doesn't just ask "What happened before?" but "How does this fit my holistic experience of the world?" Logic with Trick: This is the most fascinating part. The "Trick-based reasoning" as a way to handle social complexity and adversarial environments. In Multi-Agent Systems (MAS), if an agent only follows rigid logic, it’s easily manipulated. "Tricks" are the non-traditional heuristics that allow for strategic flexibility. Architectural Bottlenecks: We are still fighting the von Neumann bottleneck. It suggests that for AI to actually reason at scale, we need architectures that mimic the brain's parallel processing of "experience units" rather than just sequential data blocks. The Practitioner’s Take: If you’re building agents, don’t just give them a database. Give them an "experience base." We need to move from "Knowledge-Based" to "Experience-Driven" systems if we want to move past the chatbot era. #FutureOfAI #ArtificialIntelligence #TechTrends #MultiAgentSystems #CognitiveComputing

19

1 Comment

💸 Todd Chrisley hacked the system by walking right in. He pulled off a $30 million fraud campaign against several major banks using fake tax returns, doctored bank statements, and charisma as a toolkit. Victims: Synovus Bank (primary victim), Regions Bank, Fidelity Bank, Truist Bank (formerly SunTrust) Court records show Synovus issued significant loans based on paperwork that would’ve failed a middle-school science project. Chrisley’s toolkit was theatrical, not technical, and psychologically effective. The psychological firewalls were breached. He breached trust. 🛡️ Cybersecurity & Risk Failures from the Banks’ Side: Human-facing system failures. Social Engineering. 1️⃣ Failure to Perform Third-Party Document Verification He submitted fake bank statements, credit reports, and tax returns. No one cross-referenced with the IRS, source banks, or accountants. 🧠 Cyber equivalent: Accepting unsigned, unauthenticated code in a production system. 2️⃣ Lack of Anomaly Detection on Repeat Applications He reused the same forged documents across banks with minor edits. No system caught near-duplicates or metadata red flags. 🧠 Cyber equivalent: No flag for repeated login attempts from different IPs with identical payloads. 3️⃣ No Verification of Business Existence or Revenue Trail He claimed ownership of fake or shell companies. No calls, no registry checks, no revenue validation. 🧠 Cyber equivalent: Trusting a domain certificate without checking the root CA. 4️⃣ Overreliance on Charisma and Lifestyle as a Trust Signal His fame and polish short-circuited normal skepticism. 🧠 Cyber equivalent: Clicking a phishing link because it says “From: CEO” in the subject line. 5️⃣ Absence of Defense in Depth in High-Dollar Lending No layered checks, no second-line approval, no slow-it-down protocol. Just one team handing over millions. 🧠 Cyber equivalent: No multi-factor authentication on a million-dollar wire. . ⚠️ Their risk posture failed, because they treated the borrower like a celebrity, not a potential threat actor. Synovus and friends got socially engineered by a reality star 🔧 So how do you patch this? Here are 3 simple upgrades immediately: 1️⃣ Mandatory Third-Party Verification for High-Dollar Loans Verify with the issuing bank, the CPA, or the IRS. Don’t just accept a PDF with a signature and vibes. 2️⃣ Require Human + AI Red Team Review Let AI flag suspicious fonts, metadata gaps, and reused templates. Let a human say: “This smells off.” Together? Almost bulletproof. 3️⃣ Add a Manual “Trust Trigger” Step Before disbursing large sums of money, ask a human question: “Have we spoken to anyone real who can vouch for this?” If not. Pause. ⚠️ It was a culture failure over systems failure, of blind belief in a celebrity. 🏁 We must integrate Psychology into our Cybersecurity. #DefenseInDepth #ChrisleyCase #BankingFraud #FraudPrevention #SocialEngineering #ComplianceUpgrade #TrustButVerify #RiskManagement #PsychologyCybersecurity

7

1 Comment

🚢 𝗖𝗵𝗮𝗿𝘁𝗶𝗻𝗴 𝘁𝗵𝗲 𝗙𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗔𝗻𝗮𝗹𝘆𝘁𝗶𝗰𝘀 — 𝗖𝗵𝗶𝗰𝗮𝗴𝗼 𝗘𝗱𝗶𝘁𝗶𝗼𝗻 I recently attended the Analytics Leaders Network (ALN) event, “Charting the Future of Analytics in Chicago”, and it was truly inspiring. Takeaways I’m Reflecting On: Chicago’s Powerhouse Potential: With a $895B GDP, 400+ major HQs (33 Fortune 500), and 148 colleges within 25 miles, emerging global leader in quantum technology. Chicago has unmatched scale and diversity. Early Talent & Social Capital Matter: 300k+ CPS students, 83% graduation, and nearly 60% college enrollment show a strong pipeline. But skills alone aren’t enough, students and early professionals need networks, confidence, and mentorship. Data, Collaboration & Emerging Tech: By leveraging data across sectors, from transportation to policy making, Chicago can turn insights into meaningful action. Combined with cross-sector collaboration and breakthroughs like the Chicago Quantum Exchange (CQE), the city is poised to lead on a global stage. Chicago may have missed past opportunities (biotech, early AI leadership), but I left this event bullish on Chicago and especially about its next generation. My key takeaway: When we combine data with storytelling, make the benefits tangible, and build genuine partnerships, we unlock a better future for everyone. Big thanks to Tony Lopykinski for organizing and to Tom McCoy, Dominique McKoy, Nuwan Samaraweera, and Alice L. for a thoughtful, inspiring discussion. And on a personal note, it was wonderful to catch up with both old and new friends — and special thanks to Harmeet S. Grewal Savant Labs Mike Alaimo, SHRM-SCP Mavent Analytics , and Sigma for sponsoring this fantastic event. #ChartingTheFuture #ChicagoInnovation #AnalyticsLeaders Question for you 👉 : How can we make data a positive, trusted partner in decision-making whether in business, government, or education?

17

3 Comments