Moving from Usage Metrics to Trust Loops

🎬 Episode 10 - 𝗕𝗲𝘆𝗼𝗻𝗱 𝘁𝗵𝗲 𝗣𝗿𝗼𝗺𝗽𝘁: 𝗠𝗲𝗮𝘀𝘂𝗿𝗶𝗻𝗴 "𝗛𝘂𝗺𝗮𝗻 𝗥𝗢𝗜" We spent the last two episodes diagnosing the trust crisis AI is creating. Now, it's time to fix the dashboards. 🛠️ Last week in Episode 9, we uncovered the dark side of forcing AI adoption: a massive spike in "Cultural Debt," where 80% of workers fear their peers are simply faking productivity. The problem isn't just the technology; it's our analytics. If your HR dashboards are only tracking "tools logged into" or "prompts per week," you are measuring machine efficiency, not human impact. Activity does not equal value. In today’s episode, we are looking at the antidote. To rebuild trust, People Analytics teams 📊 need to pivot from tracking usage to measuring the "Human ROI" of AI: 1️⃣ 𝗕𝘂𝗿𝗻𝗼𝘂𝘁 𝗠𝗲𝘁𝗿𝗶𝗰𝘀: Is AI actually reducing after-hours work and burnout, or is it just cramming 12 hours of output into an 8-hour pressure cooker? Measure well-being, not just output. 2️⃣ 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝗼𝗻 𝗠𝗲𝘁𝗿𝗶𝗰𝘀: Are people still talking to each other? Use Organizational Network Analysis (ONA) to ensure teams aren't retreating into isolated 'AI silos' and breaking human-to-human collaboration.  3️⃣ 𝗨𝗽𝘀𝗸𝗶𝗹𝗹𝗶𝗻𝗴 𝗠𝗲𝘁𝗿𝗶𝗰𝘀: Are we measuring how our employees' critical thinking is improving, or just their tool proficiency? We need to track the growth of uniquely human skills." AI was supposed to take the robot out of the human. But if we don't update our People Analytics to measure trust, psychological safety, and connection, we are just turning our humans into faster robots. The true ROI of AI isn't found in a server room, it's found in a thriving, trusting human workforce. 👇 HR and Analytics leaders: Are your current dashboards measuring human well-being, or just AI usage?  Dave Ulrich #PeopleAnalytics #FutureOfWork

A customer-success leader at a 1,000-person enterprise told us this week: "$80M in projected savings? That's monopoly money." She wasn't being dismissive. She was telling us why the standard B2B value-selling playbook breaks at the top of the market. Here's the pattern we keep seeing: The AE runs a great pre-sales business case. 2,000 users, 25% time savings, conservative assumptions, defensible math. The number that comes out the other end is enormous — and it's real — but past a certain scale, the size of the number stops persuading and starts numbing. "Monopoly money," she said. "It doesn't have gravity anymore." What works instead - at scale - is something different: 1. Fewer, sharper outcomes. Not 14 use cases. Three or four the customer actually wants to be measured on. 2. A shared value plan, not a slide. A living artifact that pre-sales hands to post-sales, and post-sales updates with the customer's own SMEs every quarter. 3. Anecdotal-but-honest measurement. You don't run a 300-person survey to measure error frequency. You ask three team leads, you extrapolate, and you say so out loud. 4. Non-monetary value, on equal footing. Data quality, compliance, structured-data percentage. At the eight-figure-spend level, those move budgets more than dollars do. The QBR most CSMs run today is "here are your usage metrics." No anchoring value statement, no through-line to what the customer actually bought us for. Then renewal season comes and we wonder why expansion is hard. The fix isn't more dashboards. It's a forcing function, a structure that makes "what was the original promise, and how are we tracking against it?" the first question on every call. We're building Minoa to make trust a measurable artifact - not a vibe.

Let's be real, the secret to Agentic AI working well in businesses is building trust, making sure things are super reliable, and using good systems engineering; it's all about a strong base for these smart agents. Here’s the uncomfortable math: agents fail exponentially. A 10-step workflow at 95% per-step accuracy delivers ~60% end-to-end reliability. That’s not “pretty good.” That’s unshippable for anything that touches money, customers, or compliance. And the worst failures are invisible: - Infinite loops that burn tokens like a financial denial-of-service attack - Silent failures where the API call “succeeds” but the business outcome is wrong - Hallucinated parameters that pass monitoring while breaking reality - Write actions that turn a tiny mistake into a big blast radius The fix is not “better prompting.” It’s an Architecture of Trust: treat agents like unreliable components and wrap them in deterministic framework. Minimum Viable Trust Stack (MVTS): - Strict schemas for every tool input/output - Regression suite (golden datasets) on every commit - Circuit breakers for steps, time, and cost - Incident replay to reproduce failures deterministically - OpenTelemetry traces so you can debug behavior, not vibes Then mature your operating model: - Evals that move from vibes to metrics, judges, simulations, and canaries - Observability that captures decision records and full execution traces - FinOps at span-level so runaway reasoning doesn’t become your cloud bill surprise Reality check: Hyperscalers win on governance and security. Third-party tools win on deep debugging and operational reliability. Most enterprises will land on a hybrid: Hyperscaler runtime + open telemetry piping into specialized platforms. We must stop conflating model intelligence with system reliability. The competitive advantage belongs to those who wrap probabilistic cores in deterministic frame to force business-as-usual outcomes. Build the architecture of trust, or accept that your agents will remain impressive, unscalable liabilities. If you don’t build a trust architecture, your agents aren’t assets. They’re impressive liabilities. https://lnkd.in/g7R7nvXx #AgenticAI #AIEngineering #AIOps #Observability #Evaluation #Evals #OpenTelemetry #LLMOps #AITrust #EnterpriseAI #AIProductManagement #ReliabilityEngineering #ResponsibleAI #FinOps #DigitalTransformation EXL Rohit Kapoor Vivek Jetley Vikas Bhalla Anand Logani Baljinder Singh Anita Mahon Vishal Chhibbar Narasimha Kini Gaurav Iyer Shashank Verma Vivek Vinod Karan Sood Joseph Richart Aidan McGowran Saurabh Mittal Anupam Kumar Arturo Devesa Sarika Pal Adeel J. Pankaj Khera Vikrant Saraswat Wade Olson Puneet Mehra Arun Juyal Sarat Varanasi Naval Khanna Abhay B. Mustafa Karmalawala Akhil Saraf Anurag Prakash Gupta Nabarun Sengupta

🕸️ Your Referral Strategy Isn’t Linear—It’s Organic Hospice liaisons aren’t just “doing sales.” They’re building trust ecosystems—one quiet endorsement, one connection, one belief transfer at a time. But here’s where we go wrong: We track calls and visits like they’re fuel. (They are.) But we forget to ask—how far did the car actually go? 📉 “How many calls?” 📉 “How many drop-ins?” 📉 “How many referrals?” These matter. But they only tell part of the story. Because in hospice, access isn’t just activity—it’s: ⚡ Trust velocity 🌐 Network depth 💬 How fast belief spreads—and how deep it roots. Let me show you: 🔗 Referral Chaining → Wellness Director → NP → Palliative MD → Case Manager → Weekly Referrals Each door opens the next. Because trust travels. 🌱 Network Seeding → Weekly in-services or grief groups at ALFs → You support staff—no pitch, just wt. clinics → You’re already trusted when decline comes Because you were already there. 🔄 Trust Cascade Prospecting → SNF night nurse → weekend charge → DON → IDT invite This isn’t cold access. It’s trust-based acceleration. 🕸️ Web Expansion via Nodes → MA + NP + office manager → “Who do you know across the street?” → MA group texts dialysis center Your reputation hops facilities. That’s lateral growth through shared credibility. 📣 Access Through Advocacy → You guide a daughter through crisis → She introduces you to her mom’s PCP → That physician becomes a referrer You didn’t pitch. You earned it. 📏 How Do You Measure This? ✅ Still track: calls, drop-ins, referrals, conversions. But layer in the metrics that map trust flow: 🔄 Trust Velocity (TV) 📊 • # of unsolicited intros/week • Time: first contact → 2nd-degree intro • % of referrals from downstream staff (e.g., MA → NP) 🌐 Network Depth (ND) 📊 • Avg. # of distinct roles per source • # of departments touched (SW, MD, CM, DON) • Tier 1 = 4+ nodes | Tier 2 = 2–3 | Tier 3 = shallow access 📈 Tie to Conversions • Which roles convert most consistently? • Which access points have highest ROI? • Where does trust depth = sustainable admissions? 🎯 What Does This Mean? If you’re a liaison: You’re not chasing referrals. You’re cultivating networks. You’re planting trust that blooms when the time is right. If you’re a sales leader: 📌 Stop tracking surface metrics alone. 📌 Start measuring influence density and credibility transfer. Your territory map shouldn’t look like a funnel. It should look like a living mycelial network—rooted, expanding, and nourished by relationships. Because in hospice: We don’t sell services. We transfer belief. And belief spreads faster than any script ever could. #Hospice #LiaisonLeadership #TrustBasedOutreach #ReferralChaining #NetworkSeeding #SalesPsychology #AccessThroughAdvocacy #CompassionateSales #HospiceGrowth #PolarisSupportGlobal

We don’t have a product problem. We have a behaviour problem. In FinTech, we’ve spent years building faster, shinier, “smarter” tools and yet, users keep walking away. We celebrate features shipped, dashboards filled, roadmaps delivered… But deep down, we know: That’s not progress. Because if your product launches with applause but dies in silence 30 days later. It’s not innovation. It’s the Feature Factory trap. Here’s what I’ve learned after years in this space 👇 Shift 1: Build Habits, Not Features Users don’t need another app. They need help sticking to the right behaviour. It’s easier to design one smart automated decision than to expect daily effort. Simplicity creates consistency and consistency builds trust. Shift 2: Design for Humans, Not Rational Users Behavioural Economics teaches us that people rarely act logically, they act emotionally. Our job isn’t to fix that. It’s to work with it. Use choice architecture to guide good decisions effortlessly. For example: Make a savings plan “Opt-out” instead of “Opt-in.” You’ll see engagement soar. Shift 3: Make Transparency the New Currency In finance, trust is everything. No “AI-powered” widget will matter if users don’t feel safe sharing their data. Be radically clear in communication. Simplify privacy. Use microcopy that speaks human, not legal. A calm user is a loyal user. Shift 4: Move from Output to Outcome It’s time to stop tracking how many features we launched and start measuring how many financial behaviours we improved. Replace vanity metrics (DAU, MAU) with Behaviour Change Metrics: – How many users reached their savings goal this month? – How much debt did our users actually pay off? – How did retention correlate with trust? Because the truth is — when customers win, FinTech wins. Their well-being is the business model. It’s time to shut down the Feature Factory and start building real, behaviour-driven impact. What’s the biggest behavioural barrier you’ve seen your users face? How did your product help them cross it? I’d love to hear your take. #FinTech #ProductManagement #BehavioralEconomics #ProductStrategy #Innovation #UXDesign #CustomerRetention #DigitalBanking

When we talk about evaluation in AI agents, it’s tempting to treat it like a checkbox—a set of tests you run after development, or a dashboard you look at before pushing to prod. But in reality, evaluation is something far deeper. It’s not just about checking correctness it’s about building trust in the agent’s ability to reason, adapt, and improve. The more time I spend building autonomous systems, the clearer it becomes: evaluation isn’t something you do to the agent. It’s something the agent needs to do within itself. Think of it like an internal compass—constantly asking: “Did I understand the user’s goal? Is my current plan still valid? Is this tool giving the right feedback? Am I drifting off course?” In that sense, evaluation becomes less of a static test and more of a living, embedded process that runs alongside the agent’s core loop. To get there, we need to start by instrumenting agents with evaluation hooks at every critical junction: during planning, memory updates, tool invocation, interaction with other agents, and response generation. These hooks should feed into lightweight evaluators—some rule-based, some LLM-driven, some trained on reward models—that score and log performance in context. And more importantly, we need to treat those scores as feedback, not just metrics. What’s often missing is the infrastructure to close that loop. So we have to build: 1) agent-centric logging systems that capture decision traces and context windows 2) real-time feedback routers that flag deviations and inconsistencies 3) offline simulators where agents can replay failure scenarios 4) lightweight shadow agents that propose alternative plans for comparison. This becomes even more critical when agents are working in dynamic environments—switching between tools, reasoning across long time horizons, or collaborating with other agents. They can’t just act; they need to reflect in real-time. They need to recognize when something feels off—even if the output looks superficially correct. That’s where things get interesting. We’re starting to think about evaluation not just as a QA layer, but as an integral part of the agent’s intelligence. Just like people develop instincts and checks to catch their own mistakes, agents need that too. Especially in high-stakes domains, where a small misstep can cascade into major failure. We also need to think about how agents evaluate each other, how trust and alignment work in multi-agent systems, and how to build models that know when they don’t know. All of this points toward one big idea: agents need to be eval-native. Evaluation can’t live outside the system—it has to be built into its bones. Because without that constant, self-aware feedback loop, we’re not building autonomous intelligence. We’re building something that only looks smart—until it breaks. https://lnkd.in/eRhQcDc3

A friend recently told me: “Our product is heavily used, but we still have churn issues.” I asked him: “Do your customers care about usage—or ROI?” In B2B SaaS, usage isn’t the metric that matters. ROI is. Saving time. Making money. Achieving outcomes. We say this all the time. In our sales decks. On our websites. During onboarding. But the moment users get into the product, we start measuring weird things: → Logins → Clicks → CSAT → NPS Meanwhile, the real value—the why behind the usage—is ignored. Historically, this gap was filled by Success teams. They’d set goals with customers, build plans, and follow up during onboarding or (occasionally) a QBR. But here’s the thing: QBRs are lagging. They’re generic. And they usually come too late. By the time a CSM sits down with a customer to talk about “value,” the customer has already made up their mind. Now, something is shifting. The best teams are moving beyond reactive success and toward real-time ROI. Here’s how we got there: 1. Shared plans: we stopped hiding docs. Asana, Notion, Monday made roadmaps visible. 2. Live collaboration: Slack channels replaced email. Sync got faster. Feedback loops closed. 3. Real-time ROI: Now, the best products show value as it happens. → Hours saved → Workflows automated → Revenue unlocked It’s not about usage anymore. It’s about outcomes. And it’s visible directly in the product. Or maybe inside a dashboard you share. This isn’t just better reporting. It’s how you build trust. It’s how you retain customers. It’s how you turn success into part of the product experience. Quarterly reviews can’t do that. Real-time ROI can!

One of my CSMs walked in with a forecast she was proud of. The renewal column? 98% green. Usage? 15% up for 3 straight quarters. Satisfaction surveys? Almost perfect. On paper, this account looked unshakable. Three weeks later, the renewal email landed. Subject line: “𝘛𝘩𝘢𝘯𝘬 𝘺𝘰𝘶 𝘧𝘰𝘳 𝘵𝘩𝘦 𝘱𝘢𝘳𝘵𝘯𝘦𝘳𝘴𝘩𝘪𝘱.” Translation: We are out. No warning. No request for discounts. Just a clean exit. The team was rattled. “𝘞𝘩𝘢𝘵 𝘮𝘰𝘳𝘦 𝘤𝘰𝘶𝘭𝘥 𝘸𝘦 𝘩𝘢𝘷𝘦 𝘥𝘰𝘯𝘦?” This enterprise customer had been with us for more than 2 years. That is when we realized the gap. We had built a relationship with the people using the product. But we never earned mind share with the people funding the product. Here is the uncomfortable truth: Green dashboards lull you into thinking risk is low. But executives do not see green, they see budgets. What saved us in future deals was not better usage metrics, it was building three new habits: 🔹 Speaking the language of CFOs and CIOs not just end users. 🔹 Creating multiple champions inside every account, so one departure does not sink the ship. 🔹 Making ROI the hero of every conversation not product features. When we made that shift, we stopped being “the vendor our teams like” and started being “the partner leadership needs.” That is the real insurance policy for renewals. So let me ask you, when the budget conversation happens behind closed doors, will your customer’s execs fight to keep you or cut you?

After running daily huddles with AI engineers, architects, and solution advisors for the last 4 months, here's our point of view for building AI agents at enterprise scale 🎯. ☑️ Most AI today is stuck in "suggest and wait" mode. It drafts, summarizes, recommends — then waits for a human to act. That's not transformation. Transformation starts when AI is trusted to act with guardrails. 🧑 Build for Personas, not Tasks. We are stopping building isolated agents. Every agent now maps to a real human role — BUs. Persona agents orchestrate sub agents. The question shifted from "what can this agent do?" to "what does this role need to succeed?" 🔈 Platform Maturity = AI Maturity. Versioning. Deprecation strategy. Access governance. Go-Live checklists. These aren't infrastructure chores. They are the AI strategy. Without them, every production surprise is your own fault. 🖲️ Data Quality > Model Quality. The biggest blocker isn't GPT-4 vs GPT-5. It's unstructured, inaccurate, or inaccessible enterprise data. Fix the data, AI follows. We launched a data accuracy initiative because we learned this the hard way. 🤝 Trust is a KPI, not a Feature. If your agent can't earn 90%+ confidence, it doesn't touch critical workflows. We track adoption by usage frequency, # of repetable users equals to trust scores, and active feedback loops. 🙋 Human-in-the-Loop is essential, not optional. The failure rate of fully autonomous agents is still too high. Design for augmentation first. Let humans supervise outcomes, not steps. Then gradually earn the right to autonomy. The shift from 2025 to 2026 in one line: From "Can we build the AI Solution?" to "Can the org(Department/Persona) depend on AI Solution built for them?"

Make Trust Measurable ….. You know that trust is essential to good leadership—but do you have a reliable way to measure it? Instead of relying on gut instinct or vague proxies like engagement scores, you need to start measuring, tracking, and managing trust in your organization the same way you do other key metrics like financial performance or customer satisfaction. Here's how. Choose the right tool. Start by selecting a measurement model that fits your context. There are many out there; some focus on leadership behavior, others on organizational culture. The key is using a proven tool that ties specific behaviors to trust outcomes. This transforms trust from an abstract value into actionable insight. Monitor consistently. Trust isn’t static. It rises and falls based on leadership decisions, cultural dynamics, and external pressures. Just as you track performance metrics over time, you should track these trust metrics to detect early warning signs—and intervene before damage is done. Act on the data. Measurement is meaningless unless it drives action. Use trust scores to identify gaps between perception and reality, then train, coach, and adjust leadership behaviors accordingly. Benchmark externally. Finally, compare your trust metrics with other organizations in your sector. This helps you understand where you lead—or lag—and gives you a competitive edge in talent and reputation.