Navigating AI Transformation

Yesterday I had one of those rare conversations that stays with you. I sat down with Dr. Rebecca Hinds, PhD from Glean to unpack her new research, The AI Transformation 100, and it completely reframed how I think about AI in organizations. In the article I just published, I share the insights that hit me hardest — because they match exactly what I see with executives and teams every day. Here’s what you’ll learn: 🔸80% of AI transformation is just… transformation. The same human issues, politics, and leadership gaps — simply amplified by AI. 🔸AI is a megaphone. Healthy cultures accelerate. Broken workflows break faster. Rebecca’s data makes this painfully clear. 🔸Leadership behavior is the biggest adoption driver. If leaders don’t use AI themselves, their teams won’t either. 🔸100 practical ideas from 100+ leaders. Not hype — real moves happening right now. This is one of the most grounded and useful reports I’ve come across. The AI Transformation 100 releases today and I strongly encourage every business leader to read it. Access the full report and join the conversation: https://lnkd.in/e7YYwBrt And if you want the deeper story, don’t miss my full interview with Rebecca — you’ll want to watch it to the end: https://lnkd.in/egsdhVaA #GleanAmbassador #AITransformation #Leadership #ArtificialIntelligence #DigitalTransformation #ChangeManagement #FutureOfWork #BusinessStrategy #Innovation

𝗜𝗳 𝘆𝗼𝘂 𝘄𝗮𝗻𝘁 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝗮𝗻 𝗔𝗜 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆 𝗳𝗼𝗿 𝘆𝗼𝘂𝗿 𝗰𝗼𝗺𝗽𝗮𝗻𝘆, 𝘆𝗼𝘂 𝗳𝗶𝗿𝘀𝘁 𝗻𝗲𝗲𝗱 𝘁𝗼 𝗯𝘂𝗶𝗹𝗱 𝗮 𝘀𝗼𝗹𝗶𝗱 𝗱𝗮𝘁𝗮 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗮𝗻𝗱 𝗲𝗻𝗳𝗼𝗿𝗰𝗲 𝘀𝘁𝗿𝗶𝗰𝘁 𝗱𝗮𝘁𝗮 𝗵𝘆𝗴𝗶𝗲𝗻𝗲. Getting your house in order is the foundation for delivering on any AI ambition. The MIT Technology Review — based on insights from 205 C-level executives and data leaders — lays it out clearly: 𝗠𝗼𝘀𝘁 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗱𝗼 𝗻𝗼𝘁 𝗳𝗮𝗰𝗲 𝗮𝗻 𝗔𝗜 𝗽𝗿𝗼𝗯𝗹𝗲𝗺. 𝗧𝗵𝗲𝘆 𝗳𝗮𝗰𝗲 𝗰𝗵𝗮𝗹𝗹𝗲𝗻𝗴𝗲𝘀 𝗶𝗻 𝗱𝗮𝘁𝗮 𝗾𝘂𝗮𝗹𝗶𝘁𝘆, 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲, 𝗮𝗻𝗱 𝗿𝗶𝘀𝗸 𝗺𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁. Therefore, many firms are still stuck in pilots, not production. Changing that requires strong data foundations, scalable architectures, trusted partners, and a shift in how companies think about creating real value with AI. Because pilots are easy, BUT scaling AI across the enterprise is hard. 𝗛𝗲𝗿𝗲 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗸𝗲𝘆 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆𝘀: ⬇️ 1. 95% 𝗼𝗳 𝗰𝗼𝗺𝗽𝗮𝗻𝗶𝗲𝘀 𝗮𝗿𝗲 𝘂𝘀𝗶𝗻𝗴 𝗔𝗜 — 𝗯𝘂𝘁 76% 𝗮𝗿𝗲 𝘀𝘁𝘂𝗰𝗸 𝗮𝘁 𝗷𝘂𝘀𝘁 1–3 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀:   ➜ The gap between ambition and execution is huge. Scaling AI across the full business will define competitive advantage over the next 24 months. 2. 𝗗𝗮𝘁𝗮 𝗾𝘂𝗮𝗹𝗶𝘁𝘆 𝗮𝗻𝗱 𝗹𝗶𝗾𝘂𝗶𝗱𝗶𝘁𝘆 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗿𝗲𝗮𝗹 𝗯𝗼𝘁𝘁𝗹𝗲𝗻𝗲𝗰𝗸𝘀: ➜ Without curated, accessible, and trusted data, no AI strategy can succeed — no matter how powerful the models are. 3. 𝗚𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲, 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝘆, 𝗮𝗻𝗱 𝗽𝗿𝗶𝘃𝗮𝗰𝘆 𝗮𝗿𝗲 𝘀𝗹𝗼𝘄𝗶𝗻𝗴 𝗔𝗜 𝗱𝗲𝗽𝗹𝗼𝘆𝗺𝗲𝗻𝘁 — 𝗮𝗻𝗱 𝘁𝗵𝗮𝘁 𝗶𝘀 𝗮 𝗴𝗼𝗼𝗱 𝘁𝗵𝗶𝗻𝗴:   ➜ 98% of executives say they would rather be safe than first. Trust, not speed, will win in the next AI wave. 4. 𝗦𝗽𝗲𝗰𝗶𝗮𝗹𝗶𝘇𝗲𝗱, 𝗯𝘂𝘀𝗶𝗻𝗲𝘀𝘀-𝘀𝗽𝗲𝗰𝗶𝗳𝗶𝗰 𝗔𝗜 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲𝘀 𝘄𝗶𝗹𝗹 𝗱𝗿𝗶𝘃𝗲 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝘃𝗮𝗹𝘂𝗲:  ➜ Generic generative AI (chatbots, text generation) is table stakes. True differentiation will come from custom, domain-specific applications. 5. 𝗟𝗲𝗴𝗮𝗰𝘆 𝘀𝘆𝘀𝘁𝗲𝗺𝘀 𝗮𝗿𝗲 𝗮 𝗺𝗮𝗷𝗼𝗿 𝗱𝗿𝗮𝗴 𝗼𝗻 𝗔𝗜 𝗮𝗺𝗯𝗶𝘁𝗶𝗼𝗻𝘀:  ➜ Firms sitting on fragmented, outdated infrastructure are finding that retrofitting AI into legacy systems is often more costly than building new foundations. 6. 𝗖𝗼𝘀𝘁 𝗿𝗲𝗮𝗹𝗶𝘁𝗶𝗲𝘀 𝗮𝗿𝗲 𝗵𝗶𝘁𝘁𝗶𝗻𝗴 𝗵𝗮𝗿𝗱: ➜ From GPUs to energy bills, AI is not cheap — and mid-sized companies face the biggest barriers. Smart firms are building realistic ROI models that go beyond hype. 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗮 𝗳𝘂𝘁𝘂𝗿𝗲-𝗿𝗲𝗮𝗱𝘆 𝗔𝗜 𝗲𝗻𝘁𝗲𝗿𝗽𝗿𝗶𝘀𝗲 𝗶𝘀𝗻’𝘁 𝗮𝗯𝗼𝘂𝘁 𝗰𝗵𝗮𝘀𝗶𝗻𝗴 𝘁𝗵𝗲 𝗻𝗲𝘅𝘁 𝗺𝗼𝗱𝗲𝗹 𝗿𝗲𝗹𝗲𝗮𝘀𝗲.   𝗜𝘁’𝘀 𝗮𝗯𝗼𝘂𝘁 𝘀𝗼𝗹𝘃𝗶𝗻𝗴 𝘁𝗵𝗲 𝗵𝗮𝗿𝗱 𝗽𝗿𝗼𝗯𝗹𝗲𝗺𝘀 — 𝗱𝗮𝘁𝗮, 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲, 𝗴𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲, 𝗮𝗻𝗱 𝗥𝗢𝗜 — 𝘁𝗼𝗱𝗮𝘆.

Data Integration Revolution: ETL, ELT, Reverse ETL, and the AI Paradigm Shift In recents years, we've witnessed a seismic shift in how we handle data integration. Let's break down this evolution and explore where AI is taking us: 1. ETL: The Reliable Workhorse      Extract, Transform, Load - the backbone of data integration for decades. Why it's still relevant: • Critical for complex transformations and data cleansing • Essential for compliance (GDPR, CCPA) - scrubbing sensitive data pre-warehouse • Often the go-to for legacy system integration 2. ELT: The Cloud-Era Innovator Extract, Load, Transform - born from the cloud revolution. Key advantages: • Preserves data granularity - transform only what you need, when you need it • Leverages cheap cloud storage and powerful cloud compute • Enables agile analytics - transform data on-the-fly for various use cases Personal experience: Migrating a financial services data pipeline from ETL to ELT cut processing time by 60% and opened up new analytics possibilities. 3. Reverse ETL: The Insights Activator The missing link in many data strategies. Why it's game-changing: • Operationalizes data insights - pushes warehouse data to front-line tools • Enables data democracy - right data, right place, right time • Closes the analytics loop - from raw data to actionable intelligence Use case: E-commerce company using Reverse ETL to sync customer segments from their data warehouse directly to their marketing platforms, supercharging personalization. 4. AI: The Force Multiplier AI isn't just enhancing these processes; it's redefining them: • Automated data discovery and mapping • Intelligent data quality management and anomaly detection • Self-optimizing data pipelines • Predictive maintenance and capacity planning Emerging trend: AI-driven data fabric architectures that dynamically integrate and manage data across complex environments. The Pragmatic Approach: In reality, most organizations need a mix of these approaches. The key is knowing when to use each: • ETL for sensitive data and complex transformations • ELT for large-scale, cloud-based analytics • Reverse ETL for activating insights in operational systems AI should be seen as an enabler across all these processes, not a replacement. Looking Ahead: The future of data integration lies in seamless, AI-driven orchestration of these techniques, creating a unified data fabric that adapts to business needs in real-time. How are you balancing these approaches in your data stack? What challenges are you facing in adopting AI-driven data integration?

Adopting AI tools is easy. Reimagining how we work with them is the real transformation. Across many organizations, teams are being asked to “adopt AI” without the time, training or clarity they need to feel confident. When that happens, progress becomes fragmented—some people race ahead, others hesitate, and morale drops under the weight of confusion. Real AI transformation requires more than deploying technology. It demands deeper shifts that help people work differently and unlock value: → Change management to guide teams through new ways of working → Skilling to empower every employee to thrive in an AI-powered environment → Process understanding to ensure AI augments what matters most → Technology that’s usable, ethical and aligned with business goals As this Forbes article shares, the organizations that succeed will be the ones that treat AI adoption as a human journey, not just a technical one. When teams feel equipped, supported and included in shaping the path forward, that’s when AI truly delivers. What support are you giving your teams to learn and experiment with AI? https://lnkd.in/g2pXBtjm

In my Forbes article on the future of #ai and #datastorytelling (https://lnkd.in/gk4C2BWw), I included a breakdown of some of the key elements in the data storytelling process. I scored humans 👩🏼💻 and AI 🤖 on each of them. With data storytelling being a combination of art and science, the table shows there are clear strengths on each side. 🤖 AI is well-suited to the ‘science’ 🔬 aspects that are tied to the more technical and structured tasks. For example, AI can process large, complex datasets to discover potential anomalies, trends, and patterns. 👨💻 On the other hand, humans excel in the ‘art’ 🎨 aspects that depend on contextual understanding, emotional intelligence (empathy), and creativity. Often, where one side is weak, it’s a complementary strength of the other side. For example, in advanced data analysis, AI offers powerful data exploration capabilities. However, its interpretation and conclusions will be weak without adequate contextual understanding. This is where humans can offer help. In the case of data visualization, each side offers different strengths. While AI can rapidly generate data charts, humans better understand what needs to be communicated in each data scene and how best to visualize it for specific audiences. I see AI suggesting an initial set of charts for a story and then having a human customize them for context, clarity, and visual impact. As I mentioned in the article, the future of data storytelling is augmented, not automated. I’m excited to see what insights are unlocked by partnering with AI increasingly in our data storytelling. Do you agree with my scores? Are there other issues or synergies I overlooked?

AI Product Management vs AI for Product Management: Hacks and resources for you. Regardless the path you're on, you need to evolve your PM Craft. 'Evolve' being the keyword here. 𝗙𝗼𝗿 𝗔𝗜 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 (This is for the PMs working directly with AI products) – think Research PMs, Recommendations PMs, Platform PMs, and so on. You really need to get good at handling AI's unique quirks: ✨ The Probabilistic nature of AI: It's not always 0 or 1, and you've got to navigate that uncertainty. ✨ The Deep dependency on good quality data: Garbage in, garbage out. You're constantly thinking about data quality. ✨ Developing deep AI awareness: This is key but it's not about you getting too deep into technical concepts you won't need. My secret hack is to make it a habit to read research blogs from big tech companies. Google AI, Meta AI, OpenAI and attending technical conferences. Here are some: -Google AI Blog: https://ai.google/ -DeepMind's blog https://lnkd.in/g3mi8Xxy -Meta AI Blog: https://ai.meta.com/blog/ -OpenAI Research Blog: https://lnkd.in/gR_kPSkt -Microsoft AI Blog: https://lnkd.in/gYkW63yz -Amazon Science Blog: https://lnkd.in/gMJzQrGG You'll literally see what's going to be the next big product in the next two years. The original Transformers paper came out in 2017 – a PM on top of their craft could have foreseen Generative AI tools coming years ago. 𝗙𝗼𝗿 𝗔𝗜 𝗳𝗼𝗿 𝗣𝗿𝗼𝗱𝘂𝗰𝘁 𝗠𝗮𝗻𝗮𝗴𝗲𝗺𝗲𝗻𝘁 ✨ This is about leveraging AI tools to have more impact as a PM, no matter what sector you're in. It's all about adjusting your work style and experimenting to see what actually works for you. My hack here is simple but effective: train your brain to try new things. I block my calendar for 2-hour "experimentation slots." During that time, I'm creating my own tutorials, trying out new AI tools on my actual work, and following the right people. You know most of the tools by now, here are some that you might want to check out: -NotebookLM: new features getting added very often -ChatPRD: https://www.chatprd.ai/ -Productboard AI: https://lnkd.in/gm2mfeDY -ProdPad CoPilot: https://lnkd.in/gWrZZd7W -Quantilope: https://lnkd.in/g3TUJ_-9 -Dovetail: https://dovetail.com/ -Notion AI: https://lnkd.in/gfUb8yKg -Mixpanel: https://mixpanel.com/ Regardless of your seniority, being hands-on and experimenting with these tools goes a long way.

Everyone talks about agentic AI. No one shows you how to structure a production AI application from scratch. Here's the 9-layer architecture I'd follow. 1. Data Layer ↳ Ingestion pipeline (extract, clean, deduplicate, store) ↳ Chunking service (strategy depends on your content type) ↳ Embedding pipeline (batch indexing + incremental updates) ↳ Vector database with hybrid search (dense + sparse) 2. Retrieval Layer ↳ Query preprocessing (rewriting, expansion, decomposition) ↳ Hybrid retrieval (semantic + keyword) ↳ Reranking (cross-encoder second pass for precision) ↳ Source filtering (metadata, file-level, domain-level) 3. Memory and State ↳ Conversation memory (sliding window or summary) ↳ Session management ↳ Semantic cache (embed queries, serve cached answers for similar questions) 4. Routing and Classification ↳ Intent classifier (what kind of question is this) ↳ Query router (which retrieval path, which prompt template) ↳ Confidence-based fallback logic 5. Generation ↳ Prompt templates (structured per query type) ↳ Prompt registry (versioned, swappable without redeploy) ↳ Grounding rules (cite sources, handle insufficient context, abstain when needed) ↳ Streaming (real token-by-token SSE, not buffered) 6. Evaluation and Quality ↳ Golden test set (bootstrapped, grown from real failures) ↳ Offline evaluation pipeline (run on every change) ↳ Online monitoring (sampled LLM-as-judge on production traces) ↳ Document grading (system checks retrieval quality before generating) 7. Security ↳ Input validation (prompt injection detection) ↳ Retrieved content filtering (poisoning detection) ↳ Output filtering (PII, credentials, sensitive data) 8. Observability ↳ Per-stage tracing (see where each query spent time and failed) ↳ User feedback capture (linked to traces) ↳ Cost per query tracking 9. Infrastructure ↳ Backend API (async, streaming capable) ↳ Frontend (containerized separately) ↳ Docker Compose for local, cloud configs for deploy ↳ Setup scripts (environment, indexing, dependencies, smoke tests) A production AI app is not an LLM call. It's a system with data, retrieval, memory, routing, generation, evaluation, security, observability and infrastructure all working together. ____ 👋 P.S. If you want to build a system like this from scratch, on your own domain, your own data, with evaluation, security and production infrastructure baked in from the start, the Engineer's RAG Accelerator covers all 9 layers hands-on. 50+ engineers from Microsoft, Adobe, Amazon, Shopify and Visa just did exactly that. The next cohort starts in April -> [Visit my website] to register ♻️ Repost to help someone think beyond the tutorial.

29% of workers in the U.S., U.K., and Europe admit to sabotaging their company’s AI strategy. Not for the reasons you may think. The Fast Company report shows this is not about skill or technical readiness. Many employees understand the tools. The issue is rooted beneath the surface. Employees are avoiding AI, feeding it low quality inputs, or working around it because they do not trust how it will be used against them. The concerns are straightforward. Job displacement. No clarity on how outputs are evaluated. Tools introduced without context or training. People feel acted on, not included. That creates friction you will not see in a dashboard, but you will feel it in outcomes. From a leadership standpoint, this is a clear signal. When people resist, incentives and expectations are not aligned. If AI is framed as a cost reduction effort, employees will protect themselves. If success metrics are unclear, they fall back to familiar ways of working. If leaders are not explicit about how AI informs decisions, trust erodes quickly. There is also a security implication many teams are underestimating. When employees do not have clear guidance and practical education, they will find ways around the system. Shadow AI is a growing problem in most organizations, whether acknowledged or not. In agent-driven environments, this goes beyond inefficiency and into tangible security risk. What should leaders be doing about this? Here is where I’d start: 1. Remove ambiguity. Do not rely on static documents. Build clarity into how work is executed. Define where AI is used, where it is not, and which decisions remain human. 2. Make incentives explicit. If employees believe AI adoption leads to headcount reduction, resistance is a rational response. Align AI usage with growth and better outcomes, not replacement. 3. Invest in real enablement. Move beyond general training. Provide role-specific guidance that shows how AI improves the work in front of people. 4. Measure behavior, not rollout. Look at where AI is ignored, overridden, or bypassed. That is where the strategy is not landing.   What are your thoughts? Anything else I didn’t think of? Read full report here: https://lnkd.in/gE-qBbSK

For years, transformation focused on digitization. Today, we see a move toward continuous reinvention, where Ai connects strategy, operations, and talent in new ways. Yet without a strong data foundation, that vision remains out of reach.    Ai-ready data changes the equation. It brings consistency, context, and trust into every decision. It allows organizations to move from isolated insights to enterprise-wide intelligence. At scale, this creates a different kind of company, one that learns faster and acts with precision.    The organizations that will lead are those that take a disciplined approach. They define clear ownership of data. They embed governance into everyday processes. They modernize architecture to support both scale and flexibility. And they build a culture where data informs every level of decision-making.     Ai will continue to advance. That is certain. The real differentiator will be how well organizations prepare their data to keep pace.    Reinvention starts there. https://lnkd.in/gktFvYYj Accenture 

Achieving AI productivity gains usually means you have to slow down in order to speed up. Across Amazon, teams are using AI to get more done across a variety of functions, including software development. Teams that treat AI as a drop-in replacement or expect immediate gains without restructuring how they work consistently underperform. In recent months, we've been experimenting across hundreds of engineering teams and noticing where AI is delivering the most value. The largest productivity gains across the business have come from what we call frontier teams, and they usually took one of three paths: a pathfinder initiative with experts tackling a challenge, a structured sprint to execute on a well-defined plan, or an in-situ experiment splitting teams between existing approaches and AI-adapted workflows. The paths differ in structure but converge on the same insight. Teams achieved 4.5x, in some case more than 10x, productivity gains. They achieved this by reducing barriers to context for agentic workloads and increasing the surface area of work that can be done independently. Here's what I think are five ways to build an AI-native team: 1. Patience. Frontier teams that get the most productivity gains invest time in building agent context. When teams skip this step, agents keep making the same mistakes. At AWS, the Bedrock infrastructure team placed all code and documentation into a monorepo and kept the inline commentary that AI agents generated — treating it as persistent memory. 2. More patience. Push through learning curves and restructure to capture cross-functional expertise. The teams that quit this early never see the compounding acceleration that's achievable after a couple of weeks.  3. Feed agents instead of babysitting them. We saw one principal engineer ship a complete change with only 'a couple of hours of contiguous time' because the agent worked while the engineer moved between code reviews, operational support, and meetings. 4. Be very clear. Teams need to make intent explicit before code gets written. Teams that have clear context about what "done" looks like report that they handwrite only 1-2% of their code. This opens the door to push more commits per person per week. 5. "Shift testing left." Frontier teams build tooling so agents can run all integration tests locally and self-correct before code ever reaches the pipeline. The first few weeks of this process are going to feel slow. Start with a small, deliberate pilot before broadening this across your business. Take learnings and develop playbooks that your entire organization can use and build from. Frontier teams are possible for any organization, here's more on how we're building them at Amazon https://lnkd.in/gpY5UjCz