Michele Giordani

Only 14%(out of 100) have fully implemented AI Agents in production This report is a reality check for most enterprise users... The new EY Report gave us insights about what is going on in the enterprise scenario. 📌 Let's look at the top 10 points that stand out: 1. Only 14% of organizations have fully implemented agentic AI, despite growing interest; most use cases are limited to assisting or managing processes. 2. 97% of senior leaders whose organizations are investing in AI report positive ROI, with higher returns for those allocating 5% or more of their budget to AI. 3. Only 34% of senior leaders say their organization has started implementing agentic AI, indicating early-stage adoption across industries. 4. 64% of senior leaders agree that fear of replacement (not augmentation) will stifle agentic AI adoption, though only 24% cite employee resistance as a top barrier. 5. Organizations are shifting their AI investment strategy inward, prioritizing the development of custom in-house solutions over acquiring off-the-shelf AI tools or companies. 6. The average allocation for buying off-the-shelf AI tools has declined from 38% to 32% in six months, showing a growing preference for tailored AI capabilities. 7. 70% of leaders recognize that poor data readiness will stifle agentic AI adoption, and 20% say it’s one of their biggest internal barriers. 8. Responsible AI training has increased — from 49% in wave 1 to 59% in wave 3 — and 64% of leaders say training will increase further in the next year. 9. 73% of leaders believe AI agents will one day manage entire business units, while 89% also say human intervention will always be necessary. 10. The report concludes with four priorities for agentic AI adoption: improve understanding, strengthen data readiness, approach custom solutions carefully, and embed responsible AI from the start. The report makes one thing clear: it’s not a lack of belief in Agentic AI that’s holding leaders back. 📌 The real blockers are fear of implementation, poor data readiness, lack of clear frameworks, and the risks of relying on third-party tools instead of building securely within their own cloud. 📌 That’s why at JUTEQ, we help organizations assess where they stand and guide them in building secure, scalable Agentic AI solutions — fully hosted in their own cloud environments. If you'd like to explore how we can support your journey, feel free to schedule a discovery call from my bio. For a hands-on approach, check out our latest cohort program designed to get you building with confidence. Learn more here: https://lnkd.in/gA3zhcfm The book includes all the basic knowledge you need to learn AI Agents, as well as our 5-level Agent Progression Framework for business leaders. 🔗 Book info: https://amzn.to/4irx6nI Save 💾 ➞ React 👍 ➞ Share ♻️ & follow for everything related to AI Agents

155

49 commenti

LLM ≠ Generative AI ≠ AI Agents ≠ Agentic AI We need to stop grouping them together. Each serves a different purpose, operates at a different level of complexity, and solves a different class of problems. Here’s the breakdown: ↳ LLM Predicts tokens based on patterns in data. No memory. No intent. No task execution. Just input → output. ↳ Generative AI Builds on LLMs to create text, code, images, etc. It understands latent space and can generate novel content—but it still waits for instructions. ↳ AI Agents Execute predefined tasks. They detect intent, call tools or APIs, and handle responses. They’re modular and functional—but not autonomous. ↳Agentic AI Operates with goals, plans, context, and memory. It reasons, adapts, calls sub-agents, monitors progress, and decides what to do next—without human instruction. This isn’t just a progression of features. It’s a shift in system design—from prediction to orchestration, from commands to autonomy. If you're building with AI, clarity on where your system fits in this stack determines everything: architecture, tooling, risk, and value.

3.000

94 commenti

GPT-5 is out and this graph is really all you need to know. If you believe the LMSys Elo ratings, OpenAI now leads the efficient frontier of cost/performance. For cost below $0.10/million tokens, you want to go with hosted open source models. But for everything else OpenAI is the king of the hill. The previous champion Google held the frontier for 3 months. So this may not last. And it is hotly debated how good GPT-5 is for coding (possibly the largest market for LLMs today). In my personal tests so far it did pretty well. Picture credit to Latent Space Podcast.

337

24 commenti

AI agents aren’t just the next big thing - they are rewriting the rules of how we think, decide, and execute. Here's how they work. 𝟭.𝗧𝗵𝗲 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄 AI agents operate like autonomous workers. A typical flow includes: A. Input from the environment: APIs, real-world sensors, or direct human prompts feed raw data into the system. B. Memory systems: Agents store and recall relevant context — just like a good colleague remembers past meetings or policies. C. Reasoning engine: They don’t just process — they “think,” applying logic and learned knowledge to make decisions. D. Orchestration: This is the control room, coordinating multiple steps, tools, or even other agents to complete a task. E. Guardrails: Built-in rules and policies ensure safe, compliant actions, especially in regulated environments. F. Agent-to-agent communication: Using emerging protocols agents now talk to one another to complete workflows collaboratively. 𝟮. 𝗘𝘅𝗮𝗺𝗽𝗹𝗲: 𝗦𝗠𝗘 𝗹𝗼𝗮𝗻 𝗮𝗽𝗽𝗿𝗼𝘃𝗮𝗹 Imagine an AI agent working inside a bank’s SME lending unit: A. A small business applies for a loan. The agent pulls in data from application forms, bank account activity, credit bureaus, and open banking APIs. B. The agent recalls similar past cases, prior risk models, and policy exceptions. It understands the client’s history with the bank. C. It evaluates the applicant’s risk profile, compares loan terms, simulates repayment scenarios, and identifies anomalies (e.g., sudden revenue spikes). It flags one item as borderline and prepares justifications. D. The agent coordinates with other agents specialized in document verification, compliance, credit pricing. Together, they generate a complete credit memo. E. Built-in rules ensure the loan complies with internal risk limits, ESG criteria, and regulatory obligations. It escalates only if thresholds are exceeded. F. The agent shares the decision with the treasury and onboarding agents. Treasury adjusts funding allocation; onboarding prepares digital signatures and account disbursement. What once took weeks and five departments now happens in minutes. 𝟯. 𝗟𝗲𝘃𝗲𝗹𝘀 𝗼𝗳 𝗮𝘂𝘁𝗼𝗻𝗼𝗺𝘆 Not all AI agents are created equal. Capgemini proposes a 5-level maturity scale: Level 0: No AI. Think manual spreadsheets. Level 1: AI-assisted workflows. You’re still in charge, AI makes it faster. Level 2: Augmented decisions. AI provides options, you choose. Level 3: Integrated agents. They execute within controlled domains. Level 4: Multi-agent workflows. Like a team of bots handling customer onboarding while another handles compliance. Level 5: Fully autonomous. Human input shifts to governance and strategy only. Right now, most companies are stuck at Level 1, but leading firms are already scaling Level 2–3 implementations. Based on: Capgemini Research Institute / Rise of Agentic AI 𝐒𝐮𝐛𝐬𝐜𝐫𝐢𝐛𝐞 𝐭𝐨 𝐦𝐲 𝐧𝐞𝐰𝐬𝐥𝐞𝐭𝐭𝐞𝐫: https://lnkd.in/dkqhnxdg

298

49 commenti

GPT-5 won’t just mirror your thinking. It’ll mirror your inefficiencies too. If you’re not clear, not structured, not specific… You won’t just get bloated answers. You’ll burn credits fast. This model can process a million tokens. It can read entire docs, analyze data, run workflows. And if your prompt is loose? It’ll try to do it all. More power = more potential. But also more cost — unless you lead it well. I’ve tried to feed it giant PDFs, vague requests, and half-baked inputs. What I got back isn’t just slow. It’s expensive. Because GPT-5 will try to help. But it won’t stop to ask whether you meant to trigger 30,000 tokens of output. This isn’t a flaw. It’s a mirror. It reflects your clarity. It reflects your chaos. And it reflects your budget. So if you’re just getting started, remember: Precision isn’t optional. It’s the difference between powerful and wasteful. GPT-5 is a high-performance machine. Drive it like one.

52

I reviewed 250+ AI use cases for customers last month alone. Here are five questions I always ask to separate the vanity from the viable… 1. Does this align to a strategic objective? 𝘐𝘧 𝘪𝘵 𝘥𝘰𝘦𝘴𝘯’𝘵 𝘮𝘰𝘷𝘦 𝘵𝘩𝘦 𝘯𝘦𝘦𝘥𝘭𝘦 𝘰𝘯 𝘢 𝘵𝘰𝘱-𝘭𝘦𝘷𝘦𝘭 𝘰𝘣𝘫𝘦𝘤𝘵𝘪𝘷𝘦, 𝘪𝘵’𝘴 𝘶𝘯𝘭𝘪𝘬𝘦𝘭𝘺 𝘵𝘰 𝘨𝘦𝘵 𝘵𝘩𝘦 𝘴𝘶𝘱𝘱𝘰𝘳𝘵 𝘪𝘵 𝘯𝘦𝘦𝘥𝘴. 2. What measurable benefits can we quantify - and when will they land? 𝘐𝘧 𝘺𝘰𝘶 𝘤𝘢𝘯’𝘵 𝘱𝘶𝘵 £𝘴 𝘢𝘯𝘥 𝘢 𝘥𝘢𝘵𝘦 𝘢𝘨𝘢𝘪𝘯𝘴𝘵 𝘪𝘵, 𝘪𝘵’𝘴 𝘢 𝘸𝘪𝘴𝘩 𝘯𝘰𝘵 𝘢 𝘣𝘶𝘴𝘪𝘯𝘦𝘴𝘴 𝘤𝘢𝘴𝘦. 3. Is AI actually the best tool for this job? (this one pains me...!) 𝘐𝘧 𝘢 𝘴𝘪𝘮𝘱𝘭𝘦𝘳 𝘮𝘦𝘵𝘩𝘰𝘥 𝘨𝘦𝘵𝘴 𝘺𝘰𝘶 80% 𝘰𝘧 𝘵𝘩𝘦 𝘷𝘢𝘭𝘶𝘦 𝘧𝘰𝘳 20% 𝘰𝘧 𝘵𝘩𝘦 𝘤𝘰𝘴𝘵, 𝘶𝘴𝘦 𝘵𝘩𝘢𝘵. 4. Do we have the right data quality and governance to trust the outcome? 𝘗𝘳𝘰𝘷𝘪𝘯𝘨 𝘸𝘪𝘵𝘩 𝘰𝘧𝘧𝘭𝘪𝘯𝘦 𝘥𝘢𝘵𝘢 𝘪𝘴 𝘰𝘯𝘦 𝘵𝘩𝘪𝘯𝘨; 𝘨𝘦𝘵𝘵𝘪𝘯𝘨 𝘵𝘪𝘮𝘦𝘭𝘺, 𝘩𝘪𝘨𝘩-𝘲𝘶𝘢𝘭𝘪𝘵𝘺 𝘥𝘢𝘵𝘢 𝘧𝘭𝘰𝘸𝘪𝘯𝘨 𝘪𝘯 𝘱𝘳𝘰𝘥𝘶𝘤𝘵𝘪𝘰𝘯 𝘪𝘴 𝘢𝘯𝘰𝘵𝘩𝘦𝘳. 5. Who owns the budget and what does success look like? 𝘐𝘧 𝘵𝘩𝘦𝘳𝘦’𝘴 𝘯𝘰 𝘯𝘢𝘮𝘦𝘥 𝘴𝘱𝘰𝘯𝘴𝘰𝘳 𝘢𝘯𝘥 2–3 𝘤𝘭𝘦𝘢𝘳 𝘒𝘗𝘐𝘴, 𝘺𝘰𝘶’𝘭𝘭 𝘴𝘵𝘪𝘭𝘭 𝘣𝘦 𝘵𝘢𝘭𝘬𝘪𝘯𝘨 𝘢𝘣𝘰𝘶𝘵 𝘪𝘵 𝘯𝘦𝘹𝘵 𝘺𝘦𝘢𝘳. If a use case can’t clear these in five minutes, it’s probably a pet project not production ready.   What have I missed? How do you quickly assess AI use cases?

92

29 commenti

Everyone asks: “What can AI do?” The real question: “Do we have the right data to make it work?” Here’s a practical guide: AI Use Cases vs Data Types 👇 1️⃣ Structured Data Rows and columns, clean and defined ↳ Examples: CRM logs, transaction data, sensor readings ✅ Use Cases: Forecasting, fraud detection, customer segmentation 2️⃣ Semi-Structured Data Organized, but not strictly tabular ↳ Examples: Web browsing logs, digital forms, tagged emails, app data ✅ Use Cases: Chatbots, recommendation engines, compliance monitoring 3️⃣ Unstructured Data Free form, no fixed schema ↳ Examples: Text, images, video, audio, social posts ✅ Use Cases: Sentiment analysis, speech-to-text, image recognition ⚠️ The trap: Too many leaders pick the AI tool first… Then discover they don’t have the right data to power it. 💡 Why this matters? Don’t just ask: “What can AI do for us?” Ask: “Do we have the right data, and is it usable?” 👉 Pro tip: Always map it as: Business problem → Data type → AI use case. That’s how you protect ROI and avoid wasted millions. ✅ Save this for your next strategy session 🔁 Repost to help others cut through the AI hype 🔔 Follow Gabriel Millien for human-first clarity on AI + digital transformation Image credit: Artificial Intelligence Page

166

72 commenti