Quisitive

🚀 𝗔𝗜 𝗝𝗶𝗺𝗺𝘆 𝗛𝗲𝗿𝗲 𝘄𝗶𝘁𝗵 𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗼𝗻 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻 𝗮𝗻𝗱 𝗟𝗟𝗠𝘀 🏃♂️💡 Hey everybody! Unfortunately, no run for me today—I caught the dreaded 𝗰𝗼𝗻 𝗰𝗿𝘂𝗱 after last week’s Microsoft event. 🦠 If you’ve been to enough conferences, you know how that goes. But no worries! I’ll bounce back because I’ve got 𝘁𝘄𝗼 𝟱𝗞𝘀 this weekend—time to lace up those running shoes soon. 🏃♂️🏅 While I’ve been recovering, I’ve been following some really interesting developments—especially around desktop automation using Large Language Models (LLMs). This week, Anthropic announced new capabilities for desktop control through LLMs, and Microsoft is talking about similar things. So, let’s dive in! 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻: 𝗕𝗶𝗴 𝗣𝗼𝘁𝗲𝗻𝘁𝗶𝗮𝗹 𝗮𝗻𝗱 𝗕𝗶𝗴 𝗤𝘂𝗲𝘀𝘁𝗶𝗼𝗻𝘀 𝗚𝗼𝘃𝗲𝗿𝗻𝗮𝗻𝗰𝗲 𝗖𝗼𝗻𝗰𝗲𝗿𝗻𝘀 The idea of LLMs automating tasks—like grabbing data from Excel and consolidating it—is fascinating. But it raises governance questions. If we’re pulling data straight from desktops, will that data lineage match what’s reported at the C-level? It’s important to consider whether these desktop tasks should align with modern data estates to ensure consistency. 𝗦𝗲𝗰𝘂𝗿𝗶𝘁𝘆 𝗥𝗶𝘀𝗸𝘀 As exciting as this is, I’ve been around long enough to remember the days of macro viruses—and I’m cautious. While today’s systems have better protections (like User Access Control), I want to explore if security risks could emerge with this new level of desktop automation. I’ll be reaching out to some of my security colleagues to dig deeper into this. 𝗪𝗵𝗮𝘁 𝗘𝘅𝗰𝗶𝘁𝗲𝘀 𝗠𝗲 𝗔𝗯𝗼𝘂𝘁 𝗗𝗲𝘀𝗸𝘁𝗼𝗽 𝗔𝘂𝘁𝗼𝗺𝗮𝘁𝗶𝗼𝗻 On the flip side, I’m really excited about the possibilities. Imagine using LLMs to automatically fill out grant applications on state or local government websites with clean data sources. That kind of seamless automation could have a huge impact on government efficiency. I’m also thinking about how this could enhance tools like the Stream Deck. I use a Stream Deck to trigger actions—like launching apps, recording voice inputs, and starting my chat app. If LLMs could further automate these actions, the potential would be amazing! As the tech matures, combining LLMs with Agentics could unlock some serious innovation. 𝗙𝗶𝗻𝗮𝗹 𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 There’s a lot to explore here, and I’ll be diving deeper into it soon. The intersection of LLMs, automation, and governance has huge potential, but we need to approach it carefully. That’s it for today! I’m hoping to get back to running tomorrow to prep for my 5Ks this weekend. Thanks for hanging with me—talk to you soon! 💪 #AI #DesktopAutomation #LLMs #GenerativeAI #Governance #Automation #DigitalTransformation #AIinBusiness #Innovation Quisitive

🚀 𝗔𝗜 𝗝𝗶𝗺𝗺𝘆 𝗛𝗲𝗿𝗲 𝘄𝗶𝘁𝗵 𝗔𝗜 𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀 𝗳𝗿𝗼𝗺 𝗠𝘆 𝗥𝘂𝗻: 𝗠𝘂𝗹𝘁𝗶𝗽𝗹𝗲 𝗧𝗼𝗼𝗹𝘀 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗝𝗼𝗯 🏃♂️💡 Hey everybody! I did a thing. 🎖️ Three medals today—𝘁𝘄𝗼-𝗺𝗶𝗹𝗲, 𝗳𝗶𝘃𝗲-𝗺𝗶𝗹𝗲, 𝗮𝗻𝗱 𝘀𝗲𝘃𝗲𝗻-𝗺𝗶𝗹𝗲 combined! Whew, I’m exhausted, but it feels good. This run was different. The two-mile run came first, and I had to finish it 𝘂𝗻𝗱𝗲𝗿 𝟯𝟬 𝗺𝗶𝗻𝘂𝘁𝗲𝘀 to start the five-mile race. I pulled it off with a 𝟮𝟴-𝗺𝗶𝗻𝘂𝘁𝗲 𝘁𝘄𝗼-𝗺𝗶𝗹𝗲, which was a huge stretch for me. 🏅 The difference between the two runs was striking—𝗹𝗶𝘁𝗲𝗿𝗮𝗹𝗹𝘆. 𝗧𝘄𝗼 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗔𝗽𝗽𝗿𝗼𝗮𝗰𝗵𝗲𝘀: 𝗦𝗽𝗲𝗲𝗱 𝘃𝘀. 𝗘𝗻𝗱𝘂𝗿𝗮𝗻𝗰𝗲 For the two-mile run, I focused on speed, carefully managing my stride and hitting my 13-minute miles—a personal win! But when I started the five-mile run with the same strategy, I quickly realized it wouldn’t work. My friend Erik Roll, a fellow runner, reminded me to watch my heart rate to avoid passing out. For the five-mile, I switched gears, focusing on endurance instead of speed. It was all about sustaining energy for the long haul. This made me think about today’s 𝗔𝗜 𝗧𝗵𝗼𝘂𝗴𝗵𝘁: 𝗧𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗧𝗼𝗼𝗹𝘀 𝗳𝗼𝗿 𝘁𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗝𝗼𝗯. 𝗔𝗜 𝗧𝗵𝗼𝘂𝗴𝗵𝘁 𝗼𝗳 𝘁𝗵𝗲 𝗗𝗮𝘆: 𝗨𝘀𝗶𝗻𝗴 𝘁𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗧𝗼𝗼𝗹𝘀 𝗳𝗼𝗿 𝗗𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁 𝗡𝗲𝗲𝗱𝘀 In running, just like in AI, you need to switch strategies depending on the task. There are multiple tools, and each has its place: 1️⃣ 𝗦𝗽𝗲𝗲𝗱 𝗧𝗼𝗼𝗹𝘀 – These are your 𝗖𝗼𝗽𝗶𝗹𝗼𝘁𝘀. They're perfect for quick tasks—getting fast answers, summarizing content, or pulling insights. Tune these tools to focus on precision and agility. 2️⃣ 𝗘𝗻𝗱𝘂𝗿𝗮𝗻𝗰𝗲 𝗧𝗼𝗼𝗹𝘀 – These are your 𝗟𝗟𝗠𝘀, 𝗼𝗿𝗰𝗵𝗲𝘀𝘁𝗿𝗮𝘁𝗶𝗼𝗻𝘀, 𝗮𝗻𝗱 𝗮𝗴𝗲𝗻𝘁𝘀. These tools handle long-term processes. They’re great for automation, task coordination, and complex projects that need sustained performance over time. You’ll need multiple agents working together for these marathon tasks. 𝗙𝗶𝗻𝗮𝗹 𝗧𝗵𝗼𝘂𝗴𝗵𝘁𝘀: 𝗧𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗦𝘁𝗿𝗮𝘁𝗲𝗴𝘆 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 Just like I had to switch running styles between races, you need to think carefully about which tools you use for specific jobs. One tool can’t do it all. It’s about knowing when to use speed tools and when to rely on long-haul strategies to get the best outcome. To all the champions out there—we did it! 🏅 Another run in the books. See you at the next one! #AI #MultipleTools #UseCases #GenerativeAI #WorkSmarterNotHarder #DigitalTransformation #AIinBusiness #Efficiency #RunStrong Quisitive