Implementation Of Frameworks

📢 If you work in payments, your LinkedIn feed has probably been FLOODED with posts about Visa’s new VAMP rules. What Visa’s new #VAMP rules mean for every merchant Visa is merging its fraud and chargeback monitoring into one program: the Visa Acquirer Monitoring Program (VAMP). Simpler, yes. Easier, no. Starting April 2025, merchants must keep total disputes (fraud and non-fraud) below 1.5%. By January 2026, the threshold drops to 0.9%. Acquirers face even stricter rules, needing to maintain dispute rates under 0.5%. A major change involves Rapid Dispute Resolution (RDR). Previously, RDR cases didn't count against dispute ratios. Now, under VAMP, they do. Merchants lose a vital tool for managing chargebacks proactively. This shift impacts every merchant. Subscription businesses, digital goods providers, and high-risk sectors will face more scrutiny. Costs may rise as fines for excessive disputes become real. Merchants previously dependent on RDR must now develop alternative strategies. Clearer billing practices are essential. Robust fraud prevention tools are no longer optional. Enhanced customer communication is critical. Businesses must adapt quickly to these new requirements. Visa's message is clear: payment risk management now involves everyone. Staying compliant means actively reducing disputes. Transparency and proactive customer engagement will become the new standards. But questions remain. Will Visa’s VAMP truly enhance payment safety? Or does it simply add complexity and cost for merchants? How is your business preparing for the new rules? Do you see VAMP as beneficial or burdensome? Share your thoughts.

Why the EU Deforestation Regulation Matters for UK Retail Leaders....... If you run a UK retail business, you might have heard about the EU Deforestation Regulation (EUDR), but what’s the big deal? The EUDR is designed to stop products linked to deforestation from entering the EU market. That means if you sell or source goods like cocoa, coffee, cattle, soy, palm oil, rubber, timber - or certain products derived from them (think chocolate, leather, furniture or beef lasagnes) - you’ll need to prove they’re “deforestation-free.” And not just with a nice supplier statement. We’re talking hard evidence: geolocation data for farms, risk assessments, and a formal Due Diligence Statement submitted to EU authorities before goods cross the border. ***Why should UK retailers care?*** Even if you’re not trading in the EU, if your products touch EU supply chains, this regulation affects you. Miss the mark and you risk delays, fines, or even losing access to the EU market. For context, penalties can reach up to 4% of annual EU turnover, not something any board wants to explain. ***What’s the timetable?*** Due to ongoing changes in the EU, the timelines are still uncertain and may be pushed back. Our expectations as to the timelines: Large & medium companies: Current enforcement date is 30 Dec 2025 but we expect a six-month grace period to 30 June 2026 Micro & small enterprises: Currently 30 Jun 2026 but expect delay to Dec 2026 ***What should you be doing now?*** Here are three to check that your teams are ahead of the curve: ⭐Map your supply chain: Knowing exactly where your raw materials come from. Vague supplier declarations won’t cut it. ⭐Engaging your suppliers early: They’ll need to provide detailed data, so those conversations should be underway now. ⭐Build a compliance plan: Train your teams, set up systems for traceability, and prepare to submit Due Diligence Statements. Digital tools can make this less painful. The bigger picture? This isn’t just about ticking boxes. It’s about protecting forests, biodiversity, and the communities that depend on them. Call to Action: If your company is a BRC member, we’ve made it easier for you. You can download our Standardised Supplier Questionnaire or get involved in our Community to help your teams get ready today. 👇👇👇👇 #sustainability https://lnkd.in/ewpmAG4G

LLMs in production: 1️⃣ Lesson 1: Hallucinations Aren’t a Bug, They’re a Design Challenge We built our PCB troubleshooting system with LangChain + AutoGen, but the game-changer? Adding hallucination detection with deterministic fallbacks. The rule: When the model isn’t confident, fall back to known-good answers. Result: Trust went from 40% to 90%. 2️⃣ Lesson 2: RAG Speed > RAG Size Everyone obsesses over vector database choice. We used Pinecone + OpenAI embeddings with smart caching. The real win? Prompt templates and knowing when to retrieve vs. when to reason. → 60% faster time-to-answer→ Lower costs→ Happier users 3️⃣ Lesson 3: Generic Models = Generic Results Here’s the uncomfortable truth: Out-of-the-box ChatGPT won’t solve your specific problem. Fine-tuning with LoRA/PEFT on YOUR domain data is how you build a competitive moat. We use HuggingFace Transformers with mixed precision. The model learns your language, your context, your edge cases. 4️⃣ Lesson 4: Not Everything Belongs in the Cloud For NASA SBIR pre-work, we run real-time sensor fusion on Jetson Orin at the edge. Privacy-sensitive? Keep it local. Need to learn? Send insights to cloud. Need to respond instantly? Edge wins. The pattern: Edge-first for speed and privacy, cloud loops for continuous learning. 5️⃣ Lesson 5: Prompting IS Engineering Stop treating prompts like magic spells. Start treating them like code. → Role prompting for context→ Few-shot examples for patterns→ Chain-of-thought for reasoning→ Temperature tuning for determinism They’re your production reliability toolkit. 6️⃣ Lesson 6: Demos Impress, Architecture Delivers I’ve designed systems from concept to delivery, reviewed code across full stacks, and shipped AI to production. Here’s what separates working once from working always: → Systematic error handling→ Monitoring and observability→ Graceful degradation→ Human-in-the-loop when needed The Bottom Line: - Building AI that works in a demo takes weeks. - Building AI that works in production takes discipline. - Building AI that users trust takes both. Which lesson hit home for you? And what’s your biggest challenge putting LLMs into production? Drop a comment, I try to respond to every single one. P.S. Follow for more lessons from the intersection of AI and manufacturing.

𝗕𝘂𝗶𝗹𝗱 𝗬𝗼𝘂𝗿 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗦𝗮𝗻𝗱𝗯𝗼𝘅 Everyone wants innovation. But in large enterprises, that can morph into 𝗰𝗵𝗮𝗼𝘁𝗶𝗰 𝗽𝗶𝗹𝗼𝘁𝘀, 𝗱𝗶𝘀𝗰𝗼𝗻𝗻𝗲𝗰𝘁𝗲𝗱 𝘁𝗼𝗼𝗹𝘀, 𝗮𝗻𝗱 𝗿𝗼𝗴𝘂𝗲 𝘁𝗲𝗰𝗵 that can't scale. There’s a better way. Take a page from 𝗖𝗮𝗽𝗶𝘁𝗮𝗹 𝗢𝗻𝗲: They built an 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗹𝗮𝗯 where any team can experiment 𝘄𝗶𝘁𝗵𝗶𝗻 𝗮𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘂𝗿𝗮𝗹 𝗴𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀. ✅ Lightweight deployment ✅ Security by design ✅ A clear path to scale what works The result? • 𝗠𝗼𝗿𝗲 𝗰𝗿𝗲𝗮𝘁𝗶𝘃𝗶𝘁𝘆, 𝗹𝗲𝘀𝘀 𝗰𝗵𝗮𝗼𝘀. • 𝗠𝗼𝗿𝗲 𝗽𝗿𝗼𝗴𝗿𝗲𝘀𝘀, 𝗹𝗲𝘀𝘀 𝗿𝗲𝘄𝗼𝗿𝗸. 𝟯 𝗦𝘁𝗲𝗽𝘀 𝘁𝗼 𝗕𝘂𝗶𝗹𝗱 𝗮 𝗦𝗺𝗮𝗿𝘁𝗲𝗿 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗦𝗮𝗻𝗱𝗯𝗼𝘅 ✅ 𝟭. 𝗗𝗲𝗳𝗶𝗻𝗲 𝘁𝗵𝗲 𝗚𝘂𝗮𝗿𝗱𝗿𝗮𝗶𝗹𝘀 Innovation doesn’t mean starting from scratch. Give teams 𝗳𝗿𝗲𝗲𝗱𝗼𝗺 𝘄𝗶𝘁𝗵𝗶𝗻 𝗮 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 with security, architecture, and governance built in. ➡️ How? Pre-approved tooling, cloud templates, and compliance-ready data zones. ✅ 𝟮. 𝗢𝗽𝗲𝗻 𝘁𝗵𝗲 𝗙𝗿𝗼𝗻𝘁 𝗗𝗼𝗼𝗿 Don’t make innovation a secret handshake or Shadow IT. Make it easy for product and business teams to get started with light touch intake and support. ➡️ How? Publish clear entry criteria and offer “office hours” with architects. ✅ 𝟯. 𝗖𝗿𝗲𝗮𝘁𝗲 𝗮 𝗣𝗮𝘁𝗵 𝘁𝗼 𝗚𝗿𝗮𝗱𝘂𝗮𝘁𝗶𝗼𝗻 Don't let good ideas die in “pilot purgatory.” Define how a successful experiment scales: tech standards, funding triggers, transition steps. ➡️ How? Use decision gates based on business value, not bureaucracy. • 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗱𝗼𝗲𝘀𝗻’𝘁 𝗻𝗲𝗲𝗱 𝗰𝗵𝗮𝗼𝘀. • 𝗜𝘁 𝗻𝗲𝗲𝗱𝘀 𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝘁𝗵𝗮𝘁 𝘀𝘂𝗽𝗽𝗼𝗿𝘁𝘀 𝘀𝗽𝗲𝗲𝗱. 𝗔𝗿𝗰𝗵𝗶𝘁𝗲𝗰𝘁𝘀 𝗰𝗮𝗻 𝗹𝗲𝗮𝗱 𝗵𝗲𝗿𝗲 𝗯𝘆 𝗺𝗮𝗸𝗶𝗻𝗴 𝗶𝘁 𝘀𝗮𝗳𝗲 𝘁𝗼 𝗲𝘅𝗽𝗹𝗼𝗿𝗲 𝗮𝗻𝗱 𝗲𝗮𝘀𝘆 𝘁𝗼 𝘀𝗰𝗮𝗹𝗲. --- ➕ Follow Kevin Donovan 🔔 ♻️ Repost | 💬 Comment | 👍 Like 🚀 Join 𝐀𝐫𝐜𝐡𝐢𝐭𝐞𝐜𝐭𝐬’ 𝐇𝐮𝐛 – our newsletter & community to enhance skills, meet peers, and level-up your architecture career! Subscribe 👉 https://lnkd.in/dgmQqfu2

𝗣𝗿𝗼𝗰𝘂𝗿𝗲𝗺𝗲𝗻𝘁 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝗮 𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗼𝘂𝘀 𝗰𝘆𝗰𝗹𝗲 and definitely not an occasional burst of inspiration or isolated ideas. It’s a continuous, structured process that drives sustainable business results and operational improvements. It transforms internal requirements and external market insights into enhanced outcomes through collaboration, experimentation and adaptation. As a framework it creates an infinite cycle focusing on: 1️⃣ 𝗡𝗲𝗲𝗱𝘀 𝗜𝗱𝗲𝗻𝘁𝗶𝗳𝗶𝗰𝗮𝘁𝗶𝗼𝗻 & 𝗗𝗶𝘀𝗰𝗼𝘃𝗲𝗿𝘆 to fully understand potential areas for innovation stakeholders care about and who could be aligned with market trends and opportunities. 🔹 Innovation needs a clear purpose and pain points or opportunities to solve. 2️⃣ 𝗖𝗼𝗹𝗹𝗮𝗯𝗼𝗿𝗮𝘁𝗶𝘃𝗲 𝗖𝗼-𝗖𝗿𝗲𝗮𝘁𝗶𝗼𝗻 & 𝗜𝗱𝗲𝗮𝘁𝗶𝗼𝗻 by engaging suppliers, startups, and cross-functional teams to propose solutions and develop prototypes. 🔹 Innovation is a result of teaming up across disciplines and partnering. 3️⃣ 𝗣𝗶𝗹𝗼𝘁 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 & 𝗔𝗱𝗮𝗽𝘁𝗮𝘁𝗶𝗼𝗻 to test and continuously refine solutions in a controlled environment based on product feedback and market changes. 🔹 Innovation lives from experimenting, understanding and fine-tuning. 4️⃣ 𝗦𝗰𝗮𝗹𝗲 & 𝗗𝗲𝗹𝗶𝘃𝗲𝗿 𝗺𝗲𝗮𝘀𝘂𝗿𝗮𝗯𝗹𝗲 𝗶𝗺𝗽𝗮𝗰𝘁 that leads to the desired results such as cost efficiencies, improved workflows or enhanced supplier performance. 🔹 Innovation creates value when results can be scaled and demonstrated. 5️⃣ 𝗠𝗲𝗮𝘀𝘂𝗿𝗲 & 𝗘𝘃𝗮𝗹𝘂𝗮𝘁𝗲 𝗶𝗺𝗽𝗹𝗲𝗺𝗲𝗻𝘁𝗮𝘁𝗶𝗼𝗻 results and gather lessons learnt to inform further performance improvement or future innovation cycles. 🔹 Innovation must be measurable, both metrics and learning wise. 𝗣𝗿𝗼𝗰𝘂𝗿𝗲𝗺𝗲𝗻𝘁 𝗜𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻 𝗶𝘀 𝗮 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝗶𝗰 𝗮𝗱𝘃𝗮𝗻𝘁𝗮𝗴𝗲. One that empowers Procurement teams to move beyond transactions and enables enterprises to harness a wealth of possible outcomes resulting from a collaboration of internal experts and an ecosystem of external partners. ❓How does your organisation approach Procurement innovation, if at all? ❓Are traditional mindsets, as many commented on an earlier post, and a risk-averse culture the biggest barriers to innovate? Looking forward to read your views in the comments.

This is the guide I wish I had when I first tried to understand MCP. LLMs don’t live in isolation. They need structured tools, typed prompts, access to real data, and memory that doesn’t disappear after every reply. Most AI systems today break the moment you try to scale: → M×N integrations → brittle wrappers → duplicated logic in every agent That’s where 𝗠𝗖𝗣 (𝗠𝗼𝗱𝗲𝗹 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 𝗣𝗿𝗼𝘁𝗼𝗰𝗼𝗹) changes everything. It gives LLMs a clean, typed interface to: • Discover tools and resources • Reason with reusable prompts • Trigger actions • Delegate tasks across servers This issue breaks it all down: ✅ Core architecture (host, client, server) ✅ Message flow (init → result → shutdown) ✅ 4 server-side capabilities (resources, tools, prompts, sampling) ✅ A complete Claude + MCP Second Brain setup you can build today If you’re building modular LLM apps or multi-agent systems, you need to understand this protocol. Full breakdown inside → https://lnkd.in/gK_gYnxS ♻️ Repost to share these insights. ➕ Follow Shivani Virdi for more.

The European Commission has informed the European Parliament that it plans to delay the implementation of the #EUDR by one year until the end of 2026. 👉 "[T]he Commission is considering a postponement of the entry into application of the EUDR, currently foreseen for 30 December 2025, for one year, in order to avoid uncertainty for authorities and operational difficulties for stakeholders in the EU and third countries, and to allow time to remedy the identified risks." The Commission has acknowledged serious concerns with the IT system needed to implement the EUDR. Without adjustments, the system risks major slowdowns and disruptions that could prevent companies from registering, filing due diligence statements, or meeting customs requirements Read the letter below by Commissioner Roswall.

VAMP is coming. (What you need to know before April 1) On April 1, 2025, Visa rolls out the Visa Acquirer Monitoring Program Or VAMP VAMP introduces two metrics that trigger enforcement: • VAMP Ratio (fraud + non-fraud disputes) • Enumeration Ratio (card testing behavior) 1. Merchant VAMP Ratio (TC40 Fraud Disputes + Non-Fraud Disputes) / Total Transactions Excessive Thresholds 1.5% starting April 1, 2025 0.9% starting January 1, 2026 Applies to merchants with 1,000+ monthly disputes (fraud + non-fraud). Non-fraud disputes (TC15) with RDR will still be excluded. 2. Acquirer VAMP Ratio (All merchant TC40 + Non-Fraud Disputes) / Total Transactions Excessive 0.5% starting April 1, 2025 Above Standard 0.3% but < 0.5% starting January 1, 2026 Also requires 1,000+ monthly disputes to be classified. 3. Enumeration Ratio Enumerated Transactions / Total Transactions • Applies when 300,000+ enumerated transactions are flagged • Maximum threshold: 20% • Enumeration often signals card testing • Also requires 1,000+ monthly disputes 𝗔𝘀 𝗼𝗳 𝗩𝗶𝘀𝗮’𝘀 𝗠𝗮𝗿𝗰𝗵 𝟭𝟭 𝘂𝗽𝗱𝗮𝘁𝗲: Fraud-related disputes resolved via RDR or CDRN will still count toward your VAMP Ratio. Only non-fraud disputes (TC15) with RDR will still be excluded. If you're relying heavily on RDR or CDRN to reduce fraud dispute volume, your VAMP Ratio may be higher than expected. We see this as a wake-up call, not a reason to panic. This is your chance to build smarter, more connected fraud programs that account for both dispute outcomes and user intent. Before thresholds are enforced. 𝗔𝘀 𝗼𝗳 𝗩𝗶𝘀𝗮’𝘀 𝗠𝗮𝗿𝗰𝗵 𝟮𝟰 𝘂𝗽𝗱𝗮𝘁𝗲: There will be a 6-month Advisory Period extension. So enforcement is delayed until October 1, 2025 Your next step: • Audit your dispute and enumeration ratios • Model the impact of RDR/CDRN changes • Set real-time alerts for enumeration spikes Not in Q4. Now. You don’t need to overhaul your entire program overnight. You just need clear visibility, smart playbooks, and the flexibility to adapt fast. We’re helping Dodgeball customers get there now. ps... save this post to find it easily

🇪🇺 𝐕𝐢𝐃𝐀 𝐢𝐬 𝐛𝐫𝐢𝐧𝐠𝐢𝐧𝐠 𝐕𝐀𝐓 𝐫𝐞𝐟𝐨𝐫𝐦𝐬, 𝐚𝐧𝐝 𝐬𝐨𝐦𝐞 𝐰𝐢𝐥𝐥 𝐡𝐢𝐭 𝐬𝐨𝐨𝐧𝐞𝐫 𝐭𝐡𝐚𝐧 𝐲𝐨𝐮 t𝐡𝐢𝐧𝐤! On March 11, 2025, the EU formally adopted the VAT in the Digital Age (ViDA) proposal, aiming to streamline VAT reporting and reduce red tape for businesses across Europe. While the core ViDA measures won’t take effect until July 2028 and July 2030, some changes will come much sooner. And some VAT compliance headaches won’t disappear at all! 𝑾𝒉𝒂𝒕’𝒔 𝒄𝒉𝒂𝒏𝒈𝒊𝒏𝒈 𝒔𝒐𝒐𝒏? ✅ Mandatory e-invoicing for domestic transactions without prior EU approval ✅ Simplifications for low-revenue e-commerce sellers ✅ OSS expansion to include EV charging  𝑾𝒉𝒂𝒕 𝒘𝒊𝒍𝒍 𝒔𝒕𝒊𝒍𝒍 𝒃𝒆 𝒂 𝒄𝒉𝒂𝒍𝒍𝒆𝒏𝒈𝒆? ❌ Slow and bureaucratic VAT refunds ❌ Multiple VAT registrations needed by some businesses ❌ Country-specific VAT reporting requirements  ❌ No planned improvements to VIES VAT-ID validation  My latest Forbes article provides a breakdown of what’s changing before 2028 and what compliance issues will persist. https://lnkd.in/e6c5Di_9

There’s a lot of talk about connecting LLMs to tools, but very few teams have actually operationalized it in a way that scales. We’ve seen this up close, most early implementations break the moment you try to go beyond simple API calls or basic function routing. That’s exactly why we built an MCP server for Integration App. It gives your LLM a direct line to thousands of tools, but in a controlled, auditable, and infrastructure-friendly way. Think of it as a gateway that turns natural language into executable actions, backed by proper authentication, context isolation, rate-limiting, and observability. You don’t just connect to HubSpot, Notion, or Zendesk. You invoke composable actions that are designed to run inside your stack, with tenant-specific logic and secure data boundaries. Here’s a real example from a production use case from our friends at Trale AI: A user asks the assistant to find a contact during a meeting. A user asks an AI assistant to pull contact info. The client passes that to Integration App’s MCP server, which invokes a preconfigured HubSpot action through our workspace. It fetches the data, maps it to the model's context, and returns it straight into the UI - all in one flow, without building any of it from scratch. You can customize every layer: actions, schema, auth, execution scope. Or just use what’s already built. If you’re planning to scale your AI product into an actual operational system, not just a demo, this is the foundation you’ll want in place. It’s clean, it’s production-ready, and it lets your team stay focused on building intelligence, not plumbing. Docs, examples, and real implementation details here: https://lnkd.in/eS_Dtxbv