Branislav Šmidt💡’s Post

𝐎𝐩𝐞𝐧𝐀𝐈 𝐏𝐫𝐢𝐜𝐢𝐧𝐠 𝐌𝐨𝐝𝐞𝐥. 𝐖𝐡𝐲 𝐬𝐡𝐨𝐮𝐥𝐝 𝐲𝐨𝐮 𝐜𝐚𝐫𝐞? One of the most common pitfalls I see companies face when deploying GenAI at scale is making a good business case. Understanding how pricing models work for GenAI SaaS provider is crucial for that. But there is even a more crucial question to be asked once you understand these models: 𝐈𝐬 𝐲𝐨𝐮𝐫 𝐛𝐮𝐬𝐢𝐧𝐞𝐬𝐬 𝐦𝐨𝐝𝐞𝐥 𝐬𝐮𝐬𝐭𝐚𝐢𝐧𝐚𝐛𝐥𝐞 𝐢𝐧 𝐭𝐡𝐞 𝐞𝐫𝐚 𝐨𝐟 𝐆𝐞𝐧𝐀𝐈? This is an exercise I would recommend software and in general services companies taking very seriously, asking themselves if their model is sustainable, and 𝐦𝐨𝐫𝐞 𝐢𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐭𝐥𝐲, 𝐛𝐞𝐢𝐧𝐠 𝐚𝐛𝐥𝐞 𝐭𝐨 𝐜𝐥𝐞𝐚𝐫𝐥𝐲 𝐞𝐱𝐩𝐥𝐚𝐢𝐧 𝐭𝐡𝐞𝐢𝐫 𝐜𝐮𝐬𝐭𝐨𝐦𝐞𝐫𝐬 𝐰𝐡𝐲. Let me put you an example. OpenAI's states that for his 𝐧𝐞𝐰 𝐚𝐝𝐯𝐚𝐧𝐜𝐞𝐝 𝐯𝐨𝐢𝐜𝐞 𝐦𝐨𝐝𝐞𝐥 "Audio input costs are approximately 6¢ per minute; Audio output costs approximately 24¢ per minute". This means that companies could now "hire" AI call center digital assistants for less than 18$/ hour. That, when added to the new announcement of prompt caching will make this pricing lower in the coming months/ years. If you are a call center company, what does this mean for your business model? What is the differential value you provide your customers that makes you stronger than these pure GenAI-based offerings? Definitely food for thought for business leaders. #generativeAI #businessmodel All OpenAI pricings can be found here: https://lnkd.in/dNwHsbeJ

Working on updating our AI roadmaps, and here are some lessons learned over the last 12 months. 1. Be Flexible: Nobody can have a 12-month roadmap for their GenAI platform right now, because nobody has clear idea of what’s going to happen in this space after 12 months. 2. Be Modular: Build a set of modular components (vs monolithic service or library) and allow user to select just what they needed to accomplish the task. Do not fall in love with an LLM. 3. Define Evals: Define success criteria in term of Eval. For example, what does accuracy means for my use case and what % of accuracy is good enough to launch. 4. CTS will be 10x cheaper again: Do not feel limited by current cost to serve by LLM. This is going to be 10X cheaper in the next 12 months. 5. AI assisted fast iterations: All software products team needs to become AI first product teams. Make it easy to get and provide user feedback and learn to iterate at speed. Quarterly release cycle seems too long in this space. A monthly release cycle with feedback loop provide a competitive advantage. #AIroadmap #lessonlearned

Spent the weekend at the Microsoft Copilot Developer Camp and it’s awesome to see how much more flexible Copilot is getting. The highlight? Learning how Custom Engine Agents and Declarative Agents work: Custom Engine Agents — These give you total control. Copilot used to work more like a “black box,” but now you can get hands-on and customise it to suit your needs! You can integrate different LLMs and use developer tools like LangChain and Semantic Kernel, allowing you to build workflows that are tailored to specialised models and unique data sources. Declarative Agents — These are like the “easy button” for when you need something fast. Instead of getting into all the details, you just set what you want the agent to do (thus declarative), and it figures out the rest. Copilot is now more customisable than ever! Check https://lnkd.in/gXaRiqYD for details. Image below is from Copilot Dev Camp. #Copilot #Agents #Microsoft #AI #CopilotDevCamp

Future is vertical ! There is no moat in foundational AI models. Moat is in verticalization. And great customer experience coming from that focus. This was true for - - SaaS wave, - App revolution, - Software launches in early PC era. AI will follow the same path. Do not make for masses. Avoid going horizontal. Talk with AI researchers and build vertical products today. Remember speed is ultimate moat.

#ImportantNews: Microsoft GPT Builder feature for Copilot Pro subscribers to be discontinued. Microsoft will remove the ability to create GPTs starting July 10, 2024, and then will remove all GPTs (created by Microsoft and by customers) along with their associated GPT data also starting July 10, 2024, through July 14, 2024. Here is the link to official support page: https://lnkd.in/dqJNWHmM Why is Microsoft removing this feature- "According to Microsoft, they are continuing to evaluate their strategy for consumer Copilot extensibility and are prioritizing core product experiences, while remaining committed to developer opportunities. To this end, they are shifting their focus on GPTs to Commercial and Enterprise scenarios and are stopping GPT efforts in consumer Copilot" 🔔Follow our page to stay connected and receive updates about the latest news and developments in AI. Visit www.taritas.com to explore our AI development services. We are a gold partner with Microsoft in AI integration and development. Email: [email protected] #ainews #ai #microsoft #Copilot #CopilotPro #GPT #ainewstoday #GPTBuilder #CopilotGPT #MicrosoftCopilot #technews

After 4 weeks of testing at Air IT and a 2 hour deep dive demo of Microsoft 365 Copilot, here are my 3 takeaways. 🚀 #1 Copilot will save you time. 🌟 #2 Copilot will make you an expert. 🔮#3 There is much more to come. The first version is just the start of the AI potential in Copilot for Microsoft 365. We will be sharing the deep dive demo very soon. Thank you to Chris Williams Peter Pendlebury Lee Johnson and Paul N.. #copilotformicrosoft365 #sme #ai https://lnkd.in/eTuzm-GH

After our deep dive into the transformative capabilities of Microsoft 365 Copilot, our CCO, Spencer Lea, has just dropped his top three takeaways. 👇 #ai #copilotformicrosoft365

Another example of the insane speed in the #AI space. Microsoft came out only 3 months ago with their CoPilot GPT Builder and now they killed it, starting July 10th, deleting all existing GPTs and associated data on that date. #Microsoft obviously had challenges integrating #OpenAI's #ChatGPT Turbo into #CoPilot, as users faced huge difficulties in customizing the AI to meet their specific needs. Many users reported that, despite the tool's potential, it often pulled #data from the web rather than focusing on the uploaded sources. https://lnkd.in/dyDJkA3J

When one size doesn’t fit all, BITSUMMIT’s Customized Copilot Accelerator steps in. We tailor Microsoft Copilot to match the unique workflows of your industry, ensuring it becomes a powerful productivity tool embedded into your daily operations. Your Industry. Your Workflows. Your Copilot. Let’s build a solution that works for you. Reach out to learn more: https://lnkd.in/dBDQi8kQ #Copilot #Optimization #Productivity #MicrosoftCopilot #Healthcare #Finance #BusinessEfficiency #AI #Coding #DeveloperTools #Sales #Retail

Cost of Microsoft Copilot Studio in PAYG. This is a fact based comparison of different AI models. At the moment the cost of MS Copilot roll out across the enterprise even as PAYG is cost prohibitive. Let alone the enterprise Copilot at $538 per user p.a across an organisation of 2,000 users = $1,076,000. Show me a use case that will generate $1M in bottom line. Most organisation running at 10% netprofit will need to generate $10m in revenue to cover the cost of enterprise copilot. I haven't yet seen a usecase for this... have you? Show me where headcounts can be cut and or where additional benefit to the equivalent value can be quantified. #AI #Copilot #Microsoft #MS