Nir Shney-Dor

Two years ago, on this exact stage, we said the enterprise UI would collapse into a single text box. Tomorrow we will show what happened after that. Coding went agentic first. Every other business operation is next! The question is no longer whether agents will run your processes, it is whether your organization has the foundation to let hundreds of them run safely, or whether you end up with agent spaghetti. At Amazon Web Services (AWS) Summit Tel Aviv, Yariv Nir, CIO of Ichilov, Omri Kovalski and I are presenting Building An Agentic Hospital OS. How a cloud migration became a live agentic operating system inside Israel's largest ER hospital. 220 systems. Real clinicians. Production, not a pilot. We will also dive under the hood of AllCloud #AIFusion, the Agentic Operating System built on AWS AgentCore, and the AI Trust Zone, which enables deployment in a regulated environment. Because the point was never the agents. The point is who is managing whom. See you there. #AWSSummitTelAviv #AgenticAI #AWS #EnterpriseAI

84

‏3‏ ‏תגובות‏

Automat-it - one of the first AWS PLS Partners to take agentic AI from concept to production in live technical support operations. 37% faster actionable responses. 84% engineer adoption. 52.5% of cases helped within 30 minutes. 71.4% AI accuracy. All while growing 42% in under a year. I co-authored - designing the original 'Agentic Partner Ops' concept behind the sample solution that Automat-it used as their accelerator. But what they did with it is entirely their achievement: customised, hardened, and proven at scale. Pioneering work. https://lnkd.in/euPQVEEK Asaf Shirazi Oleg Yurchenko Dmytro Gnatko

32

‏1‏ ‏תגובה‏

The most frustrating TTM delays happen before the first line of application code is even written. That initial week of a new project, spent manually spinning up a secure cloud environment, configuring governance, and waiting on compliance sign-offs is pure, non-value-add "wait time." We see the biggest TTM wins come from automating that foundation. Implementing Infrastructure as Code (like Terraform) and managed Kubernetes turns complex scaling and deployment into an instant, automated event. This is how teams stop waiting on operations and start shipping features the moment they are ready. That is true competitive speed. https://lnkd.in/ennV-9jY

15

LLMs aren’t expensive because of tokens alone - the real cost includes infrastructure, latency tradeoffs, and a surprising amount of human and operational overhead. In this session, Michael Levinger shares practical strategies for cutting LLM deployment costs using our ATO-agent system as a case study. We’ll cover techniques like quantization, caching, prefetching, and distributed execution that reduce spend without sacrificing performance, plus real methods for estimating, monitoring, and budgeting costs so teams can scale responsibly. If you’re deploying LLMs in production (or planning to), don’t miss this one. Join us at Community Day on January 15th. https://luma.com/gr8eck5u

9

Great article on HackerNoon featuring our client Automat-it! From the article: "To address these challenges, Automat-it conducted a comprehensive FinOps audit of Hush Security’s cloud environment, combining financial accountability with engineering practices to align spending with business value. The team refined the network architecture by optimizing NAT Gateway usage and rerouting traffic through private endpoints and public IPs, reducing networking costs without affecting performance. Storage efficiency was improved by adjusting Amazon S3 retention policies, while idle resources, including 15 unused VPC endpoints and nine unattached Elastic IPs, were removed. Additionally, a proactive upgrade of Amazon EKS clusters ensured infrastructure remained supported, secure, and cost-efficient, and enrollment in Automat-it’s AWS reselling program provided immediate discounts and better payment terms." Link in the comments for full article. Dana Poleg

193

‏3‏ ‏תגובות‏

Had a call today with a VP R&D. Great team. Solid product. Growing fast. Massive AWS bill. I asked about their AI roadmap on AWS, Bedrock, AgentCore, and model usage, and his answer was: "Look, we know exactly what we want to build and how to do it. Our team is deeply expert. I asked, "Do you have enough hands and time to invest in it ?" Most of the team is on scaling product and GTM." That one sentence stuck with me. Because I hear it literally every week, from CTOs, R&D leads, DevOps, all across Israel. It's never "we don't know how." It's always "we don't have time or enough hands" That's why I think this offer totally makes sense: We Automat-it give you 40 hours of a senior AI engineer. They build one AI capability in your AWS account. You keep the code. Done. Available capabilities: → Claude Code on Bedrock → Fine-tuning on SageMaker → RAG & AI Knowledge Bases → NLP-to-SQL → Voice-to-Text Pipelines → Prompt & Context Optimization No decks. No discovery workshops. Just engineering work, delivered. If your team is too busy shipping product to start the AI stuff, that's exactly who this is for. DM me if you want in 🤙 (AWS customers, $20K+/mo spend, excl. Marketplace) #AI #AWS #IsraelTech #GenAI #Bedrock #AgentCore

55

‏29‏ ‏תגובות‏

What a week in Paris 🇫🇷✨ I was fortunate to join the AWS Startups EMEA offsite alongside my Automat-it colleagues, and our amazing OpenAI and Anthropic partners, it was packed with great meetings, epic experiences and inspiring people. A few highlights from the week 👇 ☕ Coffee chats at the barista cart that turned into real strategy sessions - with a good coffee :) 🚀 Deep discussions on how we can help startups ship AI to production, confidently and efficiently. 🏛️ Swapping meeting rooms for some of Paris’s most iconic venues 🥂 Tuesday’s Automat-it & AWS Startups Leadership Dinner at The Ritz Paris I just received this video from the dinner, and it captures the vibe perfectly 🎥 Take a look below! Huge thanks to the AWS Startups team for an incredible week, and to my Automat-it colleagues for making it happen 🙏 Now it’s time to turn these conversations into real results for the startups we work with 💪 👇 What do you think is still the hardest part for startups getting AI into production? Would love to hear your take. #AWS #AWSStartups #AWSPartners #Startups #GenAI #AIinProduction #Paris Ziv Kashtan Shlomo Perez Dana Poleg Noam Rahi Yoav Zuri Teresa Aparicio Jiménez Jason Bennett Matthew Thomson Udi Katz Nicolas Tarducci Alon Gendler Orel Strassman Nicolas Falleau Sebastian Storch Alex Riff James Daltas Marta Diaz Moran Onger Ben-Baruch 🧙‍♀️ Shay Wissotzky Cohen 🇮🇱🇬🇧 Reut Hovav Amit Marco Carmen Melero Timón Jaime Pérez Nacher Aine Cosgrove Laura Cavallini Fernando Laseca Zubeldia ☁️Ayal Fraiman Arthur Schmunk Dror Meiri Yoav Fishman

36

‏3‏ ‏תגובות‏

I used to think AWS Marketplace was just a listing. Put your product up, wait for buyers. Done. Then I started working with our GTM manager, Moran Onger Ben-Baruch 🧙‍♀️, on an early-stage SaaS startup. A few paying customers, big ambitions. 🚀 I quickly learned the listing is just the starting line. Moran's GTM bundle covers the whole path: ✅ Getting the product through AWS's own technical review (the FTR), so AWS trusts it enough to sell it ✅ Setting up the partner side properly, so AWS's sellers can actually find, recommend, and co-sell it ✅ Putting the startup in front of Automat-it's 1,000+ customers, so it gets real exposure That's how a small team gets a big sales force behind it. It's a privilege to learn beside talent like that at Automat-it. SaaS founder thinking about AWS Marketplace? DM me and I'll introduce you to Moran.

14

‏2‏ ‏תגובות‏

Hot on the heels of last week's piece on hidden cloud costs, I’m back in TechChannel for part two of my conversation with Nicole Willing. This time, we’re diving into a deeper, cultural challenge: "Keeping AWS Cloud Costs Under Control by Preventing Waste." As an AWS Community Builder and in my daily work at Automat-it, I often help teams tackle "procrastination debt": the costly industry habit of delaying optimization just to ship features faster. Preventing cloud waste isn't just about technical fixes; it requires a cultural shift: "You build it, you pay for it" – Shift bill visibility directly to the squad level. Blameless reviews – Treat sudden bill spikes like system outages so the whole team learns together. If your team is ready to tackle cloud waste and turn visibility into growth, let's connect. Ultimately, an enabled team is the best FinOps strategy you can have. You never have to fix problems that your engineers are too cost-aware to create in the first place! 🚀 Read the full follow-up here: https://lnkd.in/dMnBbWCH Amazon Web Services (AWS) AWS User Group Israel #FinOps #CloudOps FinOps Foundation

12

‏1‏ ‏תגובה‏