Trusted AI Strategy Consultant | Automating Complex Workflows with Our Agentic AI Frameworks & Canvas | Thought Leader

Joined March 2008
164 Photos and videos
Ok so I finally wrote this down... good to blog again. Is SaaS dead? Not exactly. But after 25 years in software and working with some of the biggest companies out there on AI right now, I can tell you it is definitely on borrowed time. linkedin.com/pulse/saas-dead…
1
11
Robertski retweeted
29 Jul 2025
Google is testing Opal, a vibe-coding app that allows users to create mini web apps using text prompts. This could change how non-technical users engage with app development! 1/7
1
1
1
156
Robertski retweeted
29 Jul 2025
Microsoft is testing an experimental Copilot Mode in Edge, integrating AI features to enhance user experience. This could redefine how we interact with the web! 1/8
1
1
1
129
30 Jul 2025
🛠️ 6 Ways We Make AI Agents Deterministic (Because Chaos Isn't Scalable) Every so often, I get this message: "Hey, I asked ChatGPT and it worked perfectly. Why can't the agent do that?" Short answer? Because your agent has to be deterministic. ChatGPT... isn't. AI Agent isn't ChatGPT. It uses the same API, but there's a lot more to an agent. And that's by design. LLMs are probabilistic. The same input doesn't guarantee the same output, especially if the system prompt, temperature, or context shifts even slightly. So how do we make Agentic AI systems predictable, repeatable, and reliable? Here's our go-to checklist: 🎯 1. System Instructions Set the tone, rules, and guardrails. Think of it like the agent's constitution. - Persona? Defined. - Output format? Rigid. - Off-limit topics? Hard no. - Context (like location, tools, logic)? Baked in. 📏 2. Temperature = 0 Force the model always to pick the most likely response. It's like saying: no surprises, no improvisation. Just stick to the script. 🧬 3. Fixed Seed When supported, a seed locks in pseudo-random processes so your outputs don't dance around. Think of it like AI déjà vu—on purpose. 🧠 4. Structured Prompting No open-ended "what do you think?" here. We use precise, declarative instructions and enforce format contracts (like JSON schemas or bullet-point structures) for consistent outputs. 📦 5. Controlled Environment Same model version. Same API logic. Same context. Even one change in the stack can lead to variability, so we freeze it. 📋 6. Robust Logging Every run is tracked from input to output, including model parameters and internal states where available. That's how we verify and debug determinism when things go sideways. You can't control what a user types into ChatGPT. But your agents? They're your product. And deterministic systems don't just "feel" more professional; they are more testable, auditable, and trustworthy. Got a funny client story where randomness bit you? Or a creative way you enforced structure? Please share it, I'm always collecting field notes. #AgenticAI #AIEngineering #LLMops #ChatGPT #Prompting #AIAutomation #Determinism #AITrust #AIAgents #DeterministicAI #FixedSeed #ModelTemperature #AI
1
1
98
20 Jun 2025
Most AI projects flop because no one stops to ask: “What exactly are we trying to automate?” That’s why we built this: agenticaicanvas.com/ 1/7
1
134
20 Jun 2025
Think Figma for automation design. Plan it right, share it easily, and skip the AI guesswork. 6/7
1
115
18 Jun 2025
🌟 Planning Tool for AI Agents: A Solid Start for Your AI Investment We often observed AI agent projects stalling due to inadequate planning. Therefore, we developed the Agentic AI Canvas Creator to address this issue directly. 🔗 agenticaicanvas.com/ (free to try) 📈 Business Canvas Meets Agentic AI It’s a straightforward tool to map out AI automation before you build. With the help of Mr. Data CoPilot, you can: - Start with guided prompts - Find practical wins for your business -Keep workflows focused on real goals - Eliminate fear and uncertainty from your automation plans Most AI projects trip up without precise planning. This tool helps you get it right from the start. We’re just getting started, so it’s free to try, and your feedback can help us improve. #AI #Automation
1
2
3
119
2 May 2025
🤣 “I’m too sexy for my... chatbot?” 🤦‍♂️ Look, we’re officially in the era where people are asking, “But is your AI... charming?” Not just accurate. Not just fast. Charming. That’s where we are. ChatGPT-4o is raising the bar with personality-forward design, and yes, it’s impressive. The tech is strong, the UX is engaging, and the conversational fluency is next level. But it also comes with a lesson. Just because our tools can smile doesn’t mean we stop using them responsibly. Cool tech doesn’t excuse lazy thinking. Hype doesn’t replace intent. And the AI might flirt, but we still have to keep our shirts tucked in. #AI #ChatGPT4o #ConversationalAI #TechEthics #UXDesign #AIWithPersonality #DigitalResponsibility #FutureOfAI #HumanMachineInteraction #OpenAI
2 May 2025
OpenAI's ChatGPT-4o is making waves with its advanced capabilities and unique personality. It's not just a chatbot; it's a conversational partner. 1/7
196
2 May 2025
😡 Kubert got my blood flowing with this one 🤬 Yes, AI is racing ahead, but does that mean we’re obsolete if we’re not sprinting with it every second? There has to be space to pause, think, and choose the right next step—not just the fastest.
2 May 2025
Airbnb is rolling out an AI customer service bot in the US. This move aims to enhance efficiency and provide 24/7 support. 1/7
1
119
2 May 2025
Airbnb’s approach makes sense: small bets, tested early. But if you take too long to scale, someone else eats your lunch. So where’s the line between thoughtful progress and falling behind? Are we still allowed to breathe in the era of AI, or is that a luxury now?
66
12 Apr 2025
🎙️ AI In The Middle Podcast – Coming Soon! Pat vs Rob vs Kubert, our AI referee. We’re covering strategy, AI, Agentic AI, application architecture, cloud systems, personal tools, and use cases shaped by 25 years of building, scaling, and designing solutions that get used. But there’s more. We’ll dig into the news, explore the research, and share what we’re seeing in the industry, alongside the tools we use and the questions we get asked all the time. We talk every day, and yes, it gets heated. So why not let you in on it? No fluff. No jargon. Just sharp takes, solid insights, and Kubert keeping us honest. Coming soon.
1
106
4 Apr 2025
🧠 How Agentic AI Learns: The Strategy Behind Smarter Automation Agentic AI learning methods are the backbone of systems that automate decision-making, data processing, content generation, and task coordination, all with impressive autonomy and contextual awareness. @PatrykAtTC and I have worked closely to design these learning tiers as part of our process for building Agentic AI Workflows. We always start at Tier 1 and progress upward. Each layer adds complexity, capability, and time investment. We’ve divided the learning process into four tiers. Effectively building adaptive workflows means combining some or all of these strategies, depending on your goals, complexity, and stage of maturity. 🔹 Tier 1 – Foundational Prompt Engineering and Retrieval-Augmented Generation (RAG) - This is the starting point. It defines agent behaviour and enables real-time access to knowledge. It is quick to implement and delivers immediate value. 🟢 Tier 2 – Iterative Improvement Feedback Loops and Memory with Reflection - This is where agents begin to learn. Feedback and outcomes drive refinement. The system improves with every use, gaining context and reliability. 🟣 Tier 3 – Advanced Learning Learning from Demonstration (LfD) and Reinforcement Learning - This stage supports dynamic skill-building. Agents learn through observation or interaction, ideal for complex workflows that require adaptability. 🟠 Tier 4 – Model Enhancement Fine-Tuning, Online Learning, Active Learning, and Distillation - These techniques upgrade the intelligence behind the agent. While not part of daily operations, they are essential for long-term performance and scale. ✅ The result? Agents that do more than automate. They adapt, evolve, and stay useful as your workflows grow. Check out the full breakdown with visuals and practical examples: 👇 translucentcomputing.com/blo… #AgenticAI #AIWorkflow #AIUX #LLM #AIEngineering #AIProductDesign #MachineLearning #ReinforcementLearning #PromptEngineering #RAG #WorkflowAutomation #IntelligentAutomation #AdaptiveSystems #AIAgents #LangChain #MLOps #AIInfrastructure #EnterpriseAI #AutonomousAgents #StrategicAI
2
74
27 Mar 2025
📸 Image Generation: Has OpenAI Just Taken the Lead? OpenAI's new image generation feature is incredible. Full blog: lnkd.in/gzJ5k5Hp I mean it. GPT-4o might be the most capable model out there right now, but is it the best? That's the question. We're not just talking about pretty pictures anymore. This thing renders text accurately, understands context, handles multi-turn image creation, and maintains visual consistency across prompts—all through casual, natural conversation. Do you need a photorealistic menu with handwritten labels? Sorted. A comic strip with a turbo snail? Easy. A witch deciphering Toronto's chaotic parking signage? No problem. What stands out is its practicality. This model isn't just flexing its creative muscles. It's building diagrams, infographics, UI mockups, and functional visual instructions that make sense. Impressive as it is, the space is moving fast. Others are coming. I am curious to see how long this lead holds. #OpenAI #ImageGeneration #AIArt #GenerativeAI #ArtificialIntelligence #MachineLearning #CreativeAI #DigitalArt #AIInnovation #TechArt
2
2
135