About
I believe the greatest challenge of the modern enterprise isn't a lack of data or…
Articles by Ragy
Activity
24K followers
-
Ragy Thomas posted thisThere’s an interesting contradiction happening with enterprise AI right now. Companies have more models, more use cases, and more ways to put AI to work than they did even a year ago. But the concern is still ROI and why it’s been so difficult to prove. I don’t think the problem is a lack of valuable use cases. In many cases, we’re simply making each one too expensive to build and scale. Every new application comes with its own integrations, context, governance, and infrastructure decisions. You can get the first one into production that way. You can probably get the fifth one there, too. But at some point, you have to ask whether you’re scaling AI or just scaling the amount of work required to support it. That’s where architecture goes beyond being a technical decision and becomes an economic one. The work you do to put one AI application into production should make the next one easier and less expensive. Otherwise, you’re paying the same tax over and over again. I wrote about this in Forbes Technology Council because I think we’ve been looking at AI ROI too narrowly. Of course an individual use case needs to create value. But there’s a much bigger question for enterprise leaders: are we building AI in a way that gets more expensive as we scale, or in a way that gets better? https://lnkd.in/etbtqjnB
-
Ragy Thomas reposted thisRagy Thomas reposted this"I can't think of a context in which it's not beneficial to have the goal of making the people you interact with feel just a little bit better about life for having had the interaction." 💡 This is what Dan Haley (Chief Legal Officer at Flock) said while speaking with Anna Richards, PMP, on the latest episode of In-House Outliers about the philosophy that guides how he leads his legal team and why leaving everyone a bit happier is so important. They also got into: 🧭 What it actually looks like to build the blueprint in a space with zero precedent ⚖️ Why the strength of your conviction should determine how hard you work to argue the other side 🤝 How legal stops being seen as the "no" team and starts being seen as a partner Listen to the full episode: https://lnkd.in/e8t_w7UW
-
Ragy Thomas shared thisIf every agent has its own understanding of the business, its own memory, and its own way of interpreting context, scaling agents can also scale fragmentation. The real opportunity is to make intelligence cumulative, so that what the enterprise learns in one place can improve how it operates everywhere else. This is why I think context will become one of the most important architectural questions in enterprise AI. There is already early evidence of its impact. Snowflake research cited in a recent Forbes piece found that adding a context layer to data agents improved accuracy by 20% while reducing tool calls by 39%. The numbers are compelling, and what they suggest is even more interesting: If shared context can make one agent more effective, what happens when that context becomes available across an entire enterprise? That is where I see the Enterprise Brain going. Beyond memory for individual agents toward a shared foundation where knowledge, context, and governance can compound across the organization. We’re still early, but I believe the architecture choices enterprises make now will determine how intelligent they can become later. Read more: https://loom.ly/SPbb39c
-
Ragy Thomas shared thisWe’ve spent a lot of time as an industry making AI more capable. Models can reason better and agents can take on increasingly complex work, but inside large enterprises, I don’t think adding more intelligence at the edges makes the organization more intelligent, unless that intelligence can compound. An agent that understands procurement but has no context from finance, legal, or operations is still operating inside a silo. Deploy hundreds or thousands of agents this way and we risk recreating the fragmentation of the application era at a much greater speed. The opportunity now is to build the shared architecture underneath them: memory, meaning, context, and governance that can extend across the enterprise. That is the idea behind the Enterprise Brain. I believe this will become one of the defining architectural shifts in enterprise technology. The conversation is moving quickly, and I enjoyed exploring where it could lead in a recent conversation with Forbes. Thank you, Sandy Carter, Doctor of Science (hon), for the thoughtful discussion and for bringing this conversation to a broader audience. Read the piece here: https://loom.ly/SPbb39c
-
Ragy Thomas posted thisWe celebrate successful pilots, impressive demos, and new model releases as if they’re indicators of long-term success, but they aren’t. They only tell us that AI can perform a task under the right conditions. What about when that intelligence has to operate inside the complexity of an enterprise? You won’t be able to ignore the disconnected systems, fragmented context, governance, and years of accumulated technical debt. AI exposes those problems. I think the organizations that win this next chapter will be the ones that build an architecture where intelligence can become part of how the business operates every day. It’s a shift in mindset that I included in a recent Forbes article for anyone who wants to explore the idea further. Read the full article in the comments.
-
Ragy Thomas shared thisThe challenge now is execution. Organizations are struggling because transformation doesn’t happen through a collection of AI projects. It happens when every initiative builds on a common foundation instead of creating another disconnected capability. That’s the difference between adopting AI and becoming AI Native. The organizations that move fastest over the next few years will be the ones with a clear architecture, a sequence for execution, and a roadmap that compounds with every step forward instead of restarting from scratch. The next phase of enterprise AI is about building the path, not just funding the vision.Every enterprise leader I speak with believes AI will fundamentally reshape their business. That debate is largely over.Every enterprise leader I speak with believes AI will fundamentally reshape their business. That debate is largely over.
-
Ragy Thomas posted thisThe biggest risk in enterprise AI is building a new layer of complexity every time you solve a problem. I’ve watched organizations add copilots, agents, vector databases, orchestration frameworks, and custom integrations one project at a time. Each initiative makes sense in isolation, but together, they create an AI estate that’s harder to govern, harder to evolve, and more expensive to operate. The irony is that many AI investments intended to accelerate the business end up slowing the next wave of innovation. That’s why I think we need to stop measuring AI success one deployment at a time. The way you answer if an AI initiative makes the enterprise simpler or more complex will determine whether your AI capabilities compound over time or become tomorrow’s technical debt. I wrote more about why this is happening and what organizations should do differently. Link in the comments. #AINative #EnterpriseAI #ArtificialIntelligence #DigitalTransformation #UnifyApps
-
Ragy Thomas shared thisI think we underestimate how quickly our definition of expertise changes. Driving used to mean mastering the wheel. Increasingly, it means knowing when to intervene. That’s a fundamentally different skill. The same shift is happening inside the enterprise. We’re still measuring people by how well they execute tasks, while AI is rapidly becoming better at executing many of them. The advantage moves to something else: judgment, oversight, context, and knowing when to step in. Being AI Native isn’t about replacing people with AI. It’s about recognizing that as intelligence becomes embedded into every workflow, human value shifts alongside it. The organizations that understand that shift early won’t just adopt AI faster, they’ll build entirely different operating models. The future rarely arrives all at once. It compounds, one seemingly small capability at a time. What skill do you think will become more valuable in an AI-Native world than it is today? #AINative #EnterpriseAI #ArtificialIntelligence #DigitalTransformation #UnifyAppsI think we underestimate how quickly our definition of expertise changes.I think we underestimate how quickly our definition of expertise changes.
-
Ragy Thomas posted thisEvery enterprise I talk to has AI pilots that work, yet very few have AI in production at scale. That’s because we’ve been optimizing the wrong 20%. The conversation is still centered on models. Which model? Which benchmark? Which provider? Models aren’t what’s keeping enterprises from becoming AI Native. The real bottleneck is the other 80% - the fragmented data, disconnected applications, inconsistent governance, and architectures that were never designed for intelligence to operate across the enterprise. Until we redesign the enterprise itself, every new model will produce the same result: another successful pilot that never compounds into enterprise-wide transformation. I wrote more about this in my latest Forbes article, where I introduce the 80/20 Production Gap and why the path to production starts with rethinking enterprise architecture, not choosing a better model. Read the full article in the comments below. #AINative #EnterpriseAI #ArtificialIntelligence #DigitalTransformation #UnifyApps
Ragy Thomas replied to a comment
1mo
Sandy Carter, Doctor of Science (hon) This is inevitable. Let's keep pulling the thread together
Ragy Thomas replied to a comment
1mo
Maybe the bigger shift is that companies will have to get much more explicit about how they make decisions, not just how they store knowledge. Two teams can have access to exactly the same facts and still make different calls. Capturing the reasoning behind those decisions may ultimately be as valuable as capturing the information itself.
-
Ragy Thomas liked thisRagy Thomas liked thisProud to share that Orascom Hotels Management has been awarded the AI Impact Award 2026 for pioneering a comprehensive Automation & AI transformation with UnifyApps in the Middle East. For me, the award is not just recognition of the technology we have implemented. It represents a much bigger shift in how we think about AI at Orascom Hotels Management and how Stuart Leven our CEO, is driving us from AI experiments and proofs of concept to AI native, production scale operations across our Hospitality business. I also had the opportunity to participate in a great panel discussion on “What does it really take to run AI at scale?” and share some of our experiences and lessons from the Orascom Hotels Management journey. The discussion reinforced an important point: scaling AI is not just about having better models. It requires the right architecture, data, governance, processes, people and, most importantly, a clear connection to business outcomes. It was a pleasure sharing the stage with Paul Potgieter, Dr Mahinour Ezzat, SFHEA, Maged DANIAL and Rami Bou Hamdan and learning from the perspectives they brought to the conversation. A big thank you to the Orascom Hotels Management team who are turning this vision into reality.
-
Ragy Thomas liked thisRagy Thomas liked thisI'm excited to share that I've joined the Board of Directors at Sprinklr. Over the years, I've had the opportunity to work through several technology shifts, from the evolution of traditional media to streaming, and from early AI in search to today's AI-powered chat and agentic experiences. One lesson has remained constant: technology creates the most value when it helps people better understand, engage with, and serve their customers. AI is creating an opportunity to fundamentally transform customer experience, and Sprinklr has taken an ambitious approach by building an AI-native platform designed to help enterprises unify customer engagement across channels and functions. Customer data from all channels and functions is the context layer that I believe will be a true differentiator in the agentic era. There’ll be plenty of capable models, but the real value will come from fine tuning with unique data to maximize quality at the lowest token cost for customer tasks. I'm looking forward to working with the Board and leadership team as they continue to innovate for customers and create long-term value. Grateful for the opportunity and excited for what's ahead. https://msft.it/6042aydjeSprinklr Appoints Jordi Ribas to Board of DirectorsSprinklr Appoints Jordi Ribas to Board of Directors
Experience & Education
Languages
-
Hindi
-
-
Tamil
-
-
Malayalam
-
Recommendations received
2 people have recommended Ragy
Join now to viewOther similar profiles
Explore more posts
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content