Backroom Organization Tips

Data classification and data categorization are often used interchangeably. But they are not quite the same thing. The simplest way I think about it: ➡️ Data categorization helps you describe and organize the data. ➡️ Data classification helps you assign the rules for how that data should be handled. For example, a Social Insurance Number could be categorized as: • Employee data • Customer data • Government identification data But when it comes to classification, it needs to land in one sensitivity class, such as: • High sensitivity data That distinction matters. Because categorization helps you understand what kind of data you have. Classification helps you decide what protections, access controls, handling rules, and retention requirements should apply. So while the two concepts are related, they serve different purposes. ➡️ Data categorization organizes the data. ➡️ Data classification governs the handling of the data. How would you describe them? ⚠️ Small caveat: in data science, classification can mean something different, such as predicting a label with a machine learning model. Here, I’m referring to data classification in the data governance, privacy, and security sense. Because context matters in data, too. Let's keep putting the Lights On Data! -George Firican #data #MDM #dataanalytics

Procurement: Treat suppliers as extensions of your enterprise, not transactions. Procurement Excellence | 23 NOV 2025 - In complex global markets, resilient supply chains demand partnerships built on shared destiny, not just contracts. Here are 9 Steps to Create Long-Term Supplier Partnerships: #1. Transparent Communication ↳ Co-develop comms protocols e.g. QBR ↳ Clearly share expectations, goals & challenges #2. Long-Term Contracts ↳ Replace short-term with multi year agreements. ↳ Share long-term roadmaps & cost-savings initiatives. #3. Shared Performance Metrics ↳ Jointly agree and track SMART KPIs. ↳ Define escalation paths & RCA templates #4. Early Supplier Involvement ↳ Involve and recognize vendor’s contributions. ↳ Include key suppliers in product development cycles. #5. Guarantee Timely Payments ↳ Automate payment & consider early payment discounts. ↳ Audit internal processes for bottlenecks. #6. Co-Create Innovation ↳ Create supplier ideation portals & protect IP collaboratively. ↳ Fund joint proof-of-concept projects. #7. Recognize & Reward Excellence ↳Formally acknowledge & reward outstanding suppliers. ↳Bronze (Operational Excellence), Silver (Innovation), Gold (Strategic Impact). #8. Uphold Fairness & Ethics ↳ Interactions & contractual terms are mutually beneficial. ↳ Ensure cost pressures don't force unethical labor. #9. Jointly Manage Risks ↳ Jointly identify risks & develop contingency plans. ↳ Map tier-2/3 suppliers collaboratively. In today's volatile market, Resilient supply chains are built on deep, strategic supplier partnerships. Achieving lasting, mutually beneficial supplier partnerships requires: ✅️ Deliberate strategy ✅️ Centered on trust ✅️ Shared objectives ✅️ Continuous collaboration ♻️ Repost if you find this helpful. ➕️ Follow Frederick for Procurement insights. #ProcurementExcellence #SupplierCollaboration

Personal data is highly sensitive information we entrust to internet companies, and strong regulations require these companies to handle it safely and reliably to meet security, privacy, and compliance standards. In this tech blog, Airbnb’s data science team shares how they built a data classification workflow to establish a unified strategy for identifying and classifying data across all data stores. The workflow is built on three pillars: Catalog, Detection, and Reconciliation. The Catalog pillar focuses on creating a dynamic and accurate system to identify where data resides and organize it into a comprehensive inventory. Detection addresses the question: what data might be considered personal? This step involves a detection engine structured as a pipeline to scan, validate, and control thresholds for surfacing detected results. Finally, Reconciliation ensures accurate classification by involving data owners in a human-in-the-loop process to confirm or refine detected classifications. Given the complexity of the system, the team developed metrics to assess its quality. These metrics—recall, precision, and speed—evaluate how effectively, accurately, and efficiently the classification system operates, ensuring it safeguards personal data over the long term. Additionally, the team shares strategies for governing data classification early in the process, along with best practices for improving workflows. These insights provide a clear understanding of not only the metrics but also actionable ways to enhance classification systems. Highly recommended reading for anyone interested in data governance and security. #datascience #personal #data #governance #classification #metrics – – –  Check out the "Snacks Weekly on Data Science" podcast and subscribe, where I explain in more detail the concepts discussed in this and future posts:    -- Spotify: https://lnkd.in/gKgaMvbh   -- Apple Podcast: https://lnkd.in/gj6aPBBY    -- Youtube: https://lnkd.in/gcwPeBmR https://lnkd.in/gqxuQ29E

 𝐒𝐜𝐞𝐧𝐚𝐫𝐢𝐨 : 𝐒𝐭𝐫𝐞𝐚𝐦𝐥𝐢𝐧𝐢𝐧𝐠 𝐈𝐧𝐯𝐞𝐧𝐭𝐨𝐫𝐲 𝐌𝐚𝐧𝐚𝐠𝐞𝐦𝐞𝐧𝐭 The Challenge: Our inventory management system was struggling to keep up with the growing volume of stock and sales data. The manual tracking process led to frequent stockouts and overstock situations, causing operational inefficiencies and affecting customer satisfaction. The Solution: We leveraged SQL to automate and optimize our inventory management process. Here’s how we did it: Steps: 1.Centralized Database Creation: Consolidated inventory data from multiple sources into a single SQL database.   Example Query to Create Inventory Table: CREATE TABLE Inventory ( ProductID INT PRIMARY KEY, ProductName VARCHAR(255), StockLevel INT, ReorderLevel INT, LastUpdated DATE );   2.Automated Stock Monitoring: Developed SQL queries to automatically monitor stock levels and trigger alerts for reorder points.   Example Query for Reorder Alerts: SELECT ProductID, ProductName, StockLevel FROM Inventory WHERE StockLevel <= ReorderLevel;   3.Dynamic Reporting: Created dynamic reports to track inventory levels, reorder statuses, and historical stock trends.   Example Query for Inventory Report: SELECT ProductID, ProductName, StockLevel, LastUpdated FROM Inventory ORDER BY LastUpdated DESC;   Impact: Operational Efficiency: Reduced manual tracking efforts, saving time and minimizing errors. Optimized Stock Levels: Improved inventory turnover by maintaining optimal stock levels. Enhanced Customer Satisfaction: Reduced stockouts and overstock situations, ensuring product availability.   Visuals: Include screenshots of the SQL queries, inventory reports, and a before-and-after comparison of stock levels. How do you manage inventory in your organization? Share your strategies and experiences in the comments!  follow more for Priyanka SG #SQL #InventoryManagement #DataOptimization #OperationalEfficiency #BusinessIntelligence

What if you could track every item in your inventory—without lifting a finger? AI-powered machine vision is transforming inventory management. No more manual errors, no more stock surprises. Just real-time visibility, smart forecasting, and seamless logistics integration. 🔍 Here’s how machine vision is redefining warehouse and supply chain efficiency: Smarter Decision-Making – AI provides accurate data that supports better planning and forecasting. Instant Visibility – Continuous monitoring detects stock shortages immediately. Operational Efficiency – Automated checks reduce repetitive tasks and boost productivity. Accurate Stock Data – AI eliminates manual mismatches and keeps records precise. Seamless Integration – Machine vision tools connect with ERP and logistics systems to streamline operations. From personal experience working with businesses that deal with complex inventory systems, I’ve seen how even small AI implementations can deliver measurable improvements in accuracy and time savings. Don't miss upcoming insights on Digital Transformation 🔔 Activate the bell to stay up to date! And if you want to delve deeper, take a look at the DeltalogiX blog > https://bit.ly/4hDs9HU #MachineVision #InventoryManagement #AIinBusiness

Fantastic overview from Shopify explaining how they use vision-language models for product classification. Shopify's product classification system serves as the basis for search ranking and related-product recommendations, so identifying nuance among related products is important. Shopify states that its first system used logistic regression and TF-IDF (a bag-of-words technique that scores words for importance based on frequency across a corpus and a document). This approach lacked classification depth since it was unimodal: it only used a product's description (text) as an input. Shopify determined that its classification system needed to capture more granular product details while adhering to its internal taxonomy, which spans 10,000 product categories. Using a large multimodal model for product classification enables the learning of more intricate relationships across modalities (image + text simultaneously) while also providing zero-shot classification capabilities. To build this system, Shopify implemented FP8 quantization (floating point precision reduction) to reduce memory footprint during inference, on-demand batching with NVIDIA Dynamo to classify products as they surface (versus waiting for an entire batch to fill), and key-value caching. The pipeline itself runs a two-stage prediction: one model call to yield the product's category and another (dependent on the first) to yield its attributes. The pipeline has utilized different models over time; the blog post indicates that the system currently uses Qwen2VL 7B. The pipeline processes 30MM predictions daily and has achieved an 85% classification acceptance rate from merchants. Shopify also built a system to produce training data that sources inputs from multiple LLMs, along with a custom-built model for tie-breaking. Blog post linked below.

💡The biggest gap in AI governance today? The industry is full of conceptual frameworks, standards, playbooks and guidelines that tell us what we must do, but almost nobody tells us how. What's missing is the operational framework in the middle that turns concepts into actionable paths. If you’ve ever tried to classify an AI system by risk, you probably experienced this frustration: NIST, ISO 42001, and OECD all say “classify risks.” The EU AI Act sorts systems into Prohibited, High-, Limited-, and Minimal-risk — but the categories depend almost entirely on predefined use cases. But what about the hundreds of internal tools, copilots, and emerging agentic systems that don't map cleanly? We need a universal, consistent, and explainable way to classify any AI system; a use-case-agnostic risk classification framework that is based on the first principles of #ResponsibleAI. My new article introduces a framework I developed that shifts the focus away from asking "Which regulation/use case does this fit?", or "Which risks are applicable?", to "How does this system fare along the core dimensions that matter." This approach scores 7 key factors to create a Risk Tier and sets the foundation for a proportionate controls catalogue. Why does this matter? Because today, AI systems are no longer “models.” They’re becoming behaviours with consequences. This is how we operationalize AI governance. Dive into the 7 dimensions 👇 #AIstrategy #AIGovernance #OperationalFramework #EnterpriseAI #ResponsibleAI #RAI #TrustworthyAI #AIEthics

An IT vendor manages your systems and closes tickets. A strategic partner asks what you're trying to achieve. One keeps things running. The other helps the business move forward. The difference starts with conversation. A vendor talks about uptime, tools, and tickets. A partner talks about revenue, productivity, and risk, and where the business is trying to go. They understand the business challenges before recommending solutions. This is where the best MSPs are creating more value for SMBs. That means helping SMBs: ⇥ Navigate AI adoption responsibly without exposing proprietary data. ⇥ Manage compliance as a business advantage, not a burden. ⇥ Connect technology investments to measurable business outcomes. ⇥ Make better strategic decisions with more confidence. An SMB owner navigating AI doesn't need someone to simply install another tool. They need a partner who can help manage that complexity so leadership can stay focused on growth. The good news is that many great MSPs are already built for this. If you already trust your MSP, the next step is simple: bring them deeper into the business conversation. Share where you’re trying to go, what risks you’re worried about, where the team is losing time, and what outcomes you want technology to support. The right MSP can do far more than support your systems. They can support the future of the business.

From Transactional to Transformational Partnerships “You don’t build something great by negotiating every brick, you build it by trusting the people laying them with you.” I’ve been reflecting on this in the context of vendor partnerships. Too often, we default to a mindset of control with tightly written, barbed-wire contracts designed to anticipate every possible failure mode, combined with behaviours that optimise every dollar through incremental negotiations. While governance has its place, this approach rarely leads to truly great outcomes. At best, it produces compliance. At worst, it erodes trust. The reality is that enduring success is not built on transactions, but it is built on partnerships. Partnerships require a shift in mindset. They ask us to move beyond “us and them” and instead operate as one team aligned on shared outcomes. When goals are common and success is measured in the same way, behaviour follows. Conversations become more open, problem-solving becomes more collaborative, and innovation becomes possible because both sides are invested in the same result. This is fundamentally about taking a long-term view. It means recognising that the health of the relationship matters just as much as the specifics of the agreement and trusting that when both sides are aligned and committed, the right commercial outcomes will follow. At Optus, this is the direction we are intentionally leaning into with Ericsson, Nokia, Cisco, HPE, and our broader partner ecosystem. We are working to better align our goals, define shared success metrics, and create an operating model that reinforces collaboration rather than division. This is not a one-off initiative, but the beginning of a broader shift in how we think about partnerships. Today was Ericsson’s turn — bringing both teams together to actively build that collaborative mindset and way of working. Mindset shifts of this nature don’t happen overnight. They require deliberate reinforcement, consistency in how we lead, and a collective effort to embed new behaviours across our teams. It takes time to move from protecting positions to building together, from managing contracts to truly nurturing partnerships. But that is the opportunity in front of us. To move from negotiating every brick to building something meaningful together. To move from “us and them” to “we.” To define success not as a zero-sum outcome, but as something shared. Because in the end, the strongest partnerships don’t just deliver better results, they create stronger, more resilient organisations on both sides. Ludvig LandgrenVincent HochartThaigan GovenderJustin MilfordNitin MadhusoodananKosala BandaranayakaJason WebsterRelihan MyburghPaul MilfordBelinda LoftsWajid BaryalaiMairead O'Riordan

Your vendors are bleeding you dry—not money, time. After managing 100+ vendor relationships across Microsoft, Instacart, and our portfolio companies, I built a system that cuts project timelines by 70%. The problem: You think hiring experts means abdicating responsibility. Wrong. Your vendors manage 50 other clients. You're not their priority unless you make yourself one. Four Frameworks That Actually Work: 1. Deconstruct Your Blockers Don't ask "what's the update?" Ask "what specific approval are we waiting for?" Financial? Technical? Legal? You can't fix what you can't name. I've seen 6-week delays resolved in one call once we identified the actual blocker. 2. Own the Project Management Your vendors are specialists, not coordinators. Schedule the calls. Create the docs. Connect the dots. Yes, you're doing their job. It's also the highest-leverage work you can do. 3. Demand Time Boxes "We're working on it" = infinite timeline "Engineering review takes 5-7 days" = accountability Even vague deadlines beat no deadlines. One portfolio company cut deployment cycles 60% just by requiring time estimates. 4. Confidence ≠ Commitment "We're confident about approval" isn't "It's approved." Push for binary answers. This distinction alone prevents countless surprises. The Process: Monday: Status email to all parties Wednesday: 15-min sync if blocked Friday: Document decisions + next actions Rule: Never let a week pass without documented progress Real Results: Applied this to 6 portfolio companies last quarter: Project completion: 12 weeks → 4 weeks Cost overruns: Down 40% Vendor performance: Up 70% Best part? Our vendors started using our process with other clients. Advanced Play: Create quarterly vendor scorecards. Measure response time, timeline accuracy, and technical competence. Share transparently. Performance improves within one quarter. Why This Matters: Every week of delay costs runway. Every vendor inefficiency is a competitor's opportunity. The companies that scale aren't the ones with the best vendors—they're the ones who best manage them. Your Move: Pick your worst vendor relationship. Apply one framework this week. Document what changes. Vendor management isn't sexy, but neither is running out of runway because every project takes 3x longer than it should. What vendor challenges are you facing? Share what's worked (or hasn't) below. — Enjoy this? ♻️ Repost it to your network and follow Kevin Henrikson for more. Weekly frameworks on AI, startups, leadership, and scaling. Join 2000+ subscribers today: https://lnkd.in/gstGkhJF