Ai for Business

  Over the last few days, there's been a lot of excitement around a UK court case where an AI law firm helped win a legal dispute. Most of the conversation has been about whether AI is replacing lawyers. I think we're asking the wrong question. What caught my attention wasn't just the courtroom victory. It was something much more fundamental. AI changed the starting point. 💡 Instead of beginning with a blank page, the claimant was able to prepare documents, organise evidence and build a strong foundation before expert legal representation became necessary. That changes the economics of accessing expertise. And that's a pattern we're beginning to see far beyond the legal profession. Whether it's consulting, patents, compliance, accounting or customer support, AI has the potential to make expert services more accessible by reducing the effort needed before the expert steps in. To me, that's the real story behind the headlines, and one that has implications fo...

  💡 “Haven’t we solved something like this before?” If you’ve worked in an equipment company where every order becomes a project, this question probably sounds familiar. Yet, despite years of accumulated experience, teams often find themselves recreating proposals, revisiting engineering decisions, and relearning lessons that already exist somewhere — buried in folders, reports, systems, or simply in people’s memories. As organizations grow, knowledge grows too. But the ability to reuse that knowledge efficiently often does not. In my latest article, I explore: 🔹 Why project-driven organizations struggle to reuse knowledge 🔹 The hidden cost of “forgotten expertise” 🔹 How Generative AI is changing the way companies can access and build on their accumulated experience Would love to hear your thoughts — have you seen this challenge in your organization? Further read : https://www.rajashreerajadhyax.com/post/when-every-order-is-a-project-why-is-knowledge-so-hard-to-reuse #Generativ...

  Recently, the founder of an Engineer-to-Order (ETO) biotech equipment company shared a common challenge. A custom bioreactor at a pharmaceutical plant suddenly stopped working. The field engineer was on-site, but the error was unusual. Manuals existed. Service notes were available. Yet finding the right solution took multiple calls, discussions, and document searches. 💡 His observation was simple but powerful: "The challenge is not that we don’t know enough. The challenge is finding the right knowledge quickly enough." This is a reality many field engineers face every day. ⚙️ Valuable troubleshooting knowledge is often: • Stored in people's memories rather than systems • Scattered across manuals, emails, and service records • Difficult to access when time is critical 📊 Modern equipment generates huge amounts of data, but data alone doesn't explain what's happening. Engineers still need context, past experiences, and machine-specific knowledge. 🤖 This is where...

  Moving Beyond the Draft: GenAI in Custom Manufacturing In ETO and MTO manufacturing, writing technical manuals is often a slow, manual data hunt. Engineers spend up to 30% of their time just chasing CAD models, ERP data, and supplier sheets. While some claim Generative AI can instantly "write" these manuals, the 2026 reality is much more practical. AI isn't replacing the engineer; it is acting as an automated assistant to handle the heavy lifting. ⚙️ The 70/30 Rule of AI Documentation The AI handles the 70%: It extracts data from CAD and ERP systems, builds the skeleton draft, populates part tables, and translates for global markets. The Human handles the 30%: Experienced engineers provide the critical "human-in-the-loop" expert review, adding the nuanced tribal knowledge that no software can replicate. 💡 For custom equipment manufacturers, the goal isn't to have completely AI-written manuals. The goal is to eliminate the "documentation tax" and...

In Made-to-Order (MTO) businesses, engineering teams rarely start from scratch — yet they often spend significant time recreating, modifying and searching for past designs 📐. Engineers must work across specifications, drawings, calculations, change requests and documentation while coordinating with multiple teams. Knowledge is usually scattered across files, folders and experienced individuals, making execution slower and harder to scale ⏳. AI can help by acting as an engineering copilot 🤖. It can assist in searching past designs, understanding specifications, generating documentation, analysing change impacts and helping teams reuse engineering knowledge more effectively. This can improve both speed and consistency while reducing repetitive effort 📊. In the next posts, the series will continue exploring how AI can transform other functions in MTO companies 🚀 #AIForMTO #AIInEngineering #Manufacturing #DigitalTransformation #GenerativeAI

Our CEO, Devesh Rajadhyax, recently delivered a session titled *“Beyond Search – How AI is Changing Knowledge Management”*, where he explored how Generative AI and Large Language Models are transforming the way organizations discover, manage, and interact with knowledge. The session explored the shift from traditional search-based systems to intelligent, AI-driven knowledge experiences — where users can interact conversationally with enterprise knowledge, uncover insights faster, and make better decisions across functions. As Devesh Rajadhyax noted during the talk, *“What ChatGPT did for individuals is what enterprises have been trying to achieve with Knowledge Management for decades.”* The session examined the limitations of conventional knowledge management approaches and the transformation being driven by the Generative AI era. To illustrate these changes, Devesh presented two contrasting examples — a large engineering enterprise and a small design firm.  At Cere Labs, building ...

We use apps like Zomato  and Swiggy  almost every day. Ordering food has become so easy 😌 that we rarely stop to think about what’s happening in the background ⚙️🤖 In this article, I’ve tried to break down where AI quietly works inside these apps—from what you see on the screen 📱 to how your food reaches you 🚚 It’s not very technical, just a simple way to look at something we all use 👍 If you get a few minutes ⏳, do have a read: 🔗 https://www.rajashreerajadhyax.com/post/ai-behind-the-scenes-in-swiggy---zomato If you find it useful, do share it with others as well 🤝✨ #AI  #EverydayAI #FoodDelivery  #Zomato #Swiggy #AIinBusiness #DigitalExperience  

  Vibe Working: From Tasks to Intent “Vibe Working” is one of those terms that sounds fluffy at first… until you realise it’s quietly changing how work actually gets done. We’re moving from task-first to intent-first. Instead of figuring out which tool, which menu, which steps…
you just describe what you want and iterate with AI in a back-and-forth flow. It feels faster. Lighter. More natural. But it also raises some uncomfortable questions: * If work becomes faster, how do we define value? * If AI does the building, are we getting better at reviewing? * Are communication and judgment becoming our real “hard skills”? This isn’t about hype. It’s a shift in how we think about work itself. I’ve put together a short piece breaking this down in a practical way. Here’s the link:  https://www.rajashreerajadhyax.com/post/what-exactly-is-vibe-working-the-new-flow-of-the-modern-workplace Curious—are you seeing this “vibe” in your daily work yet? #AI #FutureOfWork #GenerativeAI #Leadersh...

  Claude Mythos In what many are calling a turning point for artificial intelligence, Anthropic has introduced a new model named Claude Mythos, and it is already creating quite a buzz across the tech world. Unlike the usual AI announcements that focus on productivity or creativity, this one is different. Claude Mythos is being described as extremely powerful in cybersecurity. During testing, the model was able to find and even exploit deep software vulnerabilities, including bugs that had gone unnoticed for decades. In one case, it reportedly identified a 27-year-old flaw in a widely used system. What makes this even more interesting is that the company has decided not to release the model publicly. That is quite rare in today’s competitive AI race. According to reports, Anthropic itself is concerned about the risks. The fear is simple. If such a tool becomes widely available, it could be misused for hacking or cyberattacks. Instead, the model is being shared only with a small grou...

  What started as a small experiment with Claude CoWork ended up giving me a different perspective on AI 💡✨ Working with AI through workflows feels very different from just prompting it 🔄📊 But it also made one thing clear - without the right platform layer, this remains an individual productivity gain, not an organisational capability 🏢⚙️ Sharing a few thoughts on this and how I see this evolving 📈🚀 Here’s the link if you wish to read: 🔗 https://www.rajashreerajadhyax.com/post/claude-cowork-and-the-rise-of-ai-platforms-from-experiment-to-enterprise-reality #WorkflowAutomation #AI  #FutureOfWork #AIWorkflows #DigitalTransformation #EnterpriseAI #Innovation #TechTrends #Productivity #AIEvolution

AI is entering the courtroom — but judges are staying in control A recent study shows 60%+ judges are already using AI in their work — mainly to save time and improve efficiency. 💡 How judges are using AI: Summarizing case files & timelines instantly Suggesting questions for hearings Analyzing legal arguments Drafting initial rulings (after making decisions themselves) 🚀 Big advantage: AI helps handle heavy workloads and speeds up the judicial process. ⚠️ But there are risks: AI can make mistakes (fake citations, wrong info) Reliability is still a concern Judges must verify everything carefully 🧠 Key takeaway: AI is a support tool , not a decision-maker. Human judgment remains the most important part of justic e. 👉 The future of law = Human expertise + AI efficiency 🔗 Read full article: https://www.washingtonpost.com/nation/2026/04/02/judges-ai-hearings-rulings/ #AI #LegalTech #Innovation #FutureOfWork #ArtificialIntelligence #Law #DigitalTransformation #TechInLaw #Productivi...

  JPMorgan has introduced TradeFM, a generative AI model trained on billions of real trade events to understand market behavior as a “language.” 🔍 Instead of predicting words, it predicts trade flows enabling more realistic market simulations and deeper insights. 🚀 A major step toward AI-driven financial decision-making and smarter investing. 🔗 https://www.ai-street.co/p/jpmorgan-taught-ai-the-language-of #GenerativeAI #AIinFinance #JPMorgan #Innovation

  AI for Sales in Made-to-Order Companies 🚀 In Made-to-Order (MTO) businesses, winning an order often depends on how fast and how well a proposal is crafted. Yet, the effort behind it is rarely visible. Sales teams in these industries face unique challenges. Preparing proposals and responding to RFPs involves reading long specification documents, creating BOMs, preparing estimates and collating inputs from multiple teams — all under tight deadlines ⏳. Answering client queries requires deep knowledge of capabilities and past projects, which is often scattered across documents and individuals. Creating impactful collateral that truly reflects organizational strengths is another ongoing effort. AI can significantly ease this burden 🤖. It can assist in understanding RFPs, generating proposal drafts, and even supporting BOM creation and estimation by leveraging past data. An AI-enabled knowledge base can help teams quickly respond to client queries with accurate and relevant informati...

India is at a fascinating moment in its AI journey 🇮🇳✨ Over the past few years, most of the AI systems we’ve used have been built elsewhere 🌍. They are powerful, but not always designed for how India actually communicates. Recently, I spent some time testing Sarvam AI’s new models (30B and 105B), specifically through an India lens. Real prompts, real languages, real use cases. Here’s what stood out: ✅ Strong multilingual capability (Hindi, Marathi, and code-mixed queries) 🗣️ ✅ Clear, structured, and accessible responses 📊 ✅ Early signs of promise for India-specific use cases 🚀 But also: ⚠️ Gaps in factual accuracy, especially for policies and current affairs ⚠️ Cultural nuances still need deeper grounding ⚠️ High confidence even when unsure (a broader LLM challenge) Building AI for India is not just a smaller version of building global AI.
It comes with its own complexity including languages, dialects, code-mixing, and deeply local context. That is what makes this effort so impor...

  ⚙️Many industries operate in the Made-to-Order (MTO) model — where products or projects are designed and delivered only after receiving a customer order. This includes sectors like industrial machinery and equipment, aerospace and defence manufacturing, custom electronics, construction, shipbuilding and HVAC 🏗️🚢. While this model enables customization and strong customer value, it also brings unique challenges. Every order is different, so teams repeatedly create proposals, designs, plans and documents. Large volumes of technical information are generated across the lifecycle, but knowledge often remains scattered. Coordination across sales, engineering, procurement, manufacturing and execution becomes complex — leading to delays, estimation errors and rework. Expertise is unevenly distributed, and learning from past projects is not always systematic. This is where AI can make a meaningful difference 🤖. By helping organizations reuse knowledge and automate effort across the li...

  Why Multimodal Embeddings Could Change Enterprise Search Imagine this. A field technician sees a machine component on a factory floor, takes a photo of it, and the system instantly retrieves the relevant manual or maintenance documentation. No part numbers. No exact keywords. Just a photo. This kind of search experience is becoming possible with multimodal embeddings. In simple terms, embeddings convert things like text or images into numerical representations that capture their meaning. When two things are conceptually similar, their embeddings end up closer to each other in vector space. Recently, Google released Gemini Embedding 2, which creates embeddings for multiple types of data such as text, images, audio, and video, all in the same shared space. This means that a photo, a written description, or even a video frame of the same object can be understood as related by a search system. I ran a quick experiment using the Gemini API. I compared a sports car image with two text...