Generative AI Use Cases

NVIDIA 𝗥𝗲𝘀𝗲𝗮𝗿𝗰𝗵 𝗱𝗿𝗼𝗽𝘀 𝗻𝗲𝘄 𝗽𝗮𝗽𝗲𝗿 𝗼𝗻 “𝗦𝗺𝗮𝗹𝗹 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗙𝘂𝘁𝘂𝗿𝗲 𝗼𝗳 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗔𝗜”. ⬇️ LLMs are strong at broad, conversational tasks. 𝗕𝗨𝗧 Agentic AI is moving toward models built for narrow, repetitive functions. However, everyone is chasing bigger models. But this paper argues the exact opposite: Most agent workloads don’t need 175B params — they need precision, speed, and control. 𝗧𝗵𝗲 𝗸𝗲𝘆 𝘁𝗮𝗸𝗲𝗮𝘄𝗮𝘆? Small Language Models (SLMs) are not only good enough — they’re better for the majority of agentic use cases. 𝗛𝗲𝗿𝗲 𝗶𝘀 𝗮 𝗾𝘂𝗶𝗰𝗸 𝘀𝘂𝗺𝗺𝗮𝗿𝘆 𝗼𝗳 𝘁𝗵𝗲 𝗸𝗲𝘆 𝗽𝗼𝗶𝗻𝘁𝘀: ⬇️ 1. SLMs can already match or beat 30–70B LLMs on task-specific reasoning → From Phi-3 to DeepSeek Distill, we now have 2–9B models outperforming legacy LLMs with 10–70× faster inference. 2. Most agents just run repetitive, scoped tasks → Parsing. Routing. Tool calls. Summaries. You don’t need an all-knowing LLM — you need a fast, fine-tuned SLM that gets the job done. 3. LLMs are economically unsustainable at scale → They dominate cloud costs and energy use. SLMs offer massive savings in latency, memory, and operational overhead. 4. SLMs run on edge and consumer devices → Tools like ChatRTX show real-time agents can live on laptops or embedded systems — without phoning home to a GPU cluster. 5. Heterogeneous agent stacks are the path forward → Use LLMs sparingly for general reasoning. Let SLMs handle 80% of workflows. More modular. More efficient. More robust. 6. SLMs are easier to fine-tune and align → Lower hallucination risk, tighter output control, and better format consistency. Perfect for tool-driven agent environments. More in the comments and the paper below to download — but I’ll say this now: This paper might age like gold for every team trying to ship serious agents in production: https://lnkd.in/dx3vMQwA 𝗣.𝗦. 𝗜 𝗿𝗲𝗰𝗲𝗻𝘁𝗹𝘆 𝗹𝗮𝘂𝗻𝗰𝗵𝗲𝗱 𝗮 𝗻𝗲𝘄𝘀𝗹𝗲𝘁𝘁𝗲𝗿 𝘄𝗵𝗲𝗿𝗲 𝗜 𝘀𝗵𝗮𝗿𝗲 𝘁𝗵𝗲 𝗯𝗲𝘀𝘁 𝘄𝗲𝗲𝗸𝗹𝘆 𝗱𝗿𝗼𝗽𝘀 𝗼𝗻 𝗔𝗜 𝗮𝗴𝗲𝗻𝘁𝘀, 𝗲𝗺𝗲𝗿𝗴𝗶𝗻𝗴 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄𝘀, 𝗮𝗻𝗱 𝗵𝗼𝘄 𝘁𝗼 𝘀𝘁𝗮𝘆 𝗮𝗵𝗲𝗮𝗱 𝘄𝗵𝗶𝗹𝗲 𝗼𝘁𝗵𝗲𝗿𝘀 𝘄𝗮𝘁𝗰𝗵 𝗳𝗿𝗼𝗺 𝘁𝗵𝗲 𝘀𝗶𝗱𝗲𝗹𝗶𝗻𝗲𝘀. 𝗜𝘁’𝘀 𝗳𝗿𝗲𝗲 — 𝗮𝗻𝗱 𝗮𝗹𝗿𝗲𝗮𝗱𝘆 𝗿𝗲𝗮𝗱 𝗯𝘆 𝟮𝟬,𝟬𝟬𝟬+ 𝗽𝗲𝗼𝗽𝗹𝗲. 𝗝𝗼𝗶𝗻 𝘁𝗵𝗲𝗺 𝗵𝗲𝗿𝗲: https://lnkd.in/dbf74Y9E

Gemini, ChatGPT, Copilot… these are just some of the #generativeAI tools that we use at #MercedesBenz to create cutting-edge software features. People often ask me why we use different types of #AI rather than simplify everything and use just one. The answer is easy: no single AI model can do it all. Each tool has a specific strength and, by combining them, we can build the best possible software experience for our customers. 👍 For example, we’ve already integrated #ChatGPT through Microsoft Azure OpenAI Service into our current voice assistant system to provide up-to-date answers to knowledge-based questions. When our new #CLA arrives, we will take #genAI a step further. It will be the first Mercedes-Benz to run on #MBOS and use the new #MBUX Virtual Assistant – the next evolution of our current voice assistant. We will continue to use ChatGPT, but we can now also use Google Cloud’s new Automotive AI Agent. This is built using Google’s Gemini models so you can tap into Google Maps and ask more conversational queries about all manner of POIs. The point here is that it’s a win-win for our customers. By integrating multiple LLMs as a Multi-Agent into one intelligent voice assistant, the customer just asks a question, and the system responds using the most appropriate AI tool. The process is invisible, seamless and secure, providing quick and accurate answers every time. ➡️ How would you like AI to improve your driving experience? #Software

There’s a lot of excitement about generative AI in pharmaceutical R&D — and rightly so. But alongside the enthusiasm, there’s still real uncertainty about where & how it’s delivering value today, and what it will actually take to scale impact. A recent paper from the DISRUPT–Data Science Industry Roundtable — a cross-industry consortium I co-founded — offers a useful, data-based snapshot of the current state of play. The analysis is grounded in aggregated, anonymized benchmarking across multiple large pharma companies, which makes the insights especially practical. What I appreciate most about the paper is that it doesn’t oversell where the industry is. Where GenAI is actively being used today: ◆ Knowledge search, summarization, and synthesis across biological, chemical, and clinical data ◆ Drafting & supporting research, clinical, and regulatory-adjacent documents ◆ Assistive productivity tools for scientists, statisticians, and data teams ◆ Early traction in molecule design, though maturity & consistency remain mixed across organizations These use cases are real and, in many organizations, already moving into routine use — particularly for knowledge workflows & document generation. Where proof points are still emerging: ◆ Clear, reproducible examples of GenAI directly driving clinical-stage success ◆ Demonstrated impact beyond efficiency gains into durable biological insight ◆ Scaled, end-to-end integration into core R&D decision-making processes Much of today’s GenAI value in pharma remains assistive rather than decisive — an important & often overlooked distinction. What continues to limit scale: ◆ Data quality & availability: models amplify what they’re trained on, and biology remains noisy & incomplete ◆ Process integration: embedding GenAI into existing R&D workflows is as hard as building the models themselves ◆ Trust, robustness, and oversight: human-in-the-loop remains essential in regulated environments ◆ Organizational readiness: governance, operating models, and adoption matter as much as algorithms The takeaway for me isn’t skepticism — it’s focus. Early wins matter because they build confidence & momentum. But the step-change impact many hope for will come from doing the harder work: investing in high-quality data foundations, tightly coupling AI systems to experimental feedback, and building operating models that earn trust over time. That’s where efforts like DISRUPT are valuable — not because they provide all the answers, but because they create shared visibility. They help the industry separate real progress from noise & have more honest conversations about what it will actually take to scale impact responsibly. We’re entering a more mature phase of generative AI in medicine — one where optimism is paired with discipline, and success is measured not by what’s possible, but by what’s proven. Read more:  https://lnkd.in/e5ZZkWRE #GenerativeAI #AIinMedicine #DrugDiscovery

The use-cases for AI and GenAI are truly limitless.    One of the new ways Deloitte is leveraging #GenAI is by supporting internal audit teams in their development of #AI strategies and applied capabilities. Not only are these tools supporting teams in the day-to-day audit process, but they are allowing them to build toward future-state operating models.    Here are a few of the ways Deloitte is offering AI-powered tools for the audit process:    Dynamic Risk Assessments – We utilize AI to develop end-to-end assessment capabilities to create more proactive models, resulting in a dynamic and iterative #risk assessment lifecycle that evolves with the org’s needs.    AI-on-Demand PODs – Our AI-on-Demand Product Oriented Delivery (POD) service delivery model consists of a team of engineers and designers to help clients develop customizable AI solutions that follow our Trustworthy AI Framework ™ (https://deloi.tt/3ywy7K8).    Automated SOX Scoping – We work with our clients to utilize AI to increase efficiency and save time during the Sarbanes-Oxley (SOX) scoping process. The statistical algorithms we put into place help clients develop a more accurate and risk-aligned scope for their SOX programs.    You can read more about how AI is changing the #audit landscape, here: https://deloi.tt/4d4xRBa Chris Griffin, Trevear Thomas, Dipti Gulati, Lynne Sterrett

Do you rely on one large generalist model to power multiple use cases, or do you build a suite of specialized models fine-tuned for specific tasks? Large Language Models (LLMs) act as the generalists. One model can handle many functions across financial services: -Fraud Detection -Automated Investing -Customer Service Chatbots -Personalized Banking -Consumer Loan Underwriting -This flexibility makes them ideal for exploration, rapid prototyping, and -scenarios where breadth of understanding matters more than hyper-optimization. Small Language Models (SLMs) act as the specialists. Each is optimized for a single task, such as: -Loan Qualification -Consumer Loan Underwriting -Fraud Detection -The benefit? Efficiency, accuracy, and cost control. By narrowing the scope, SLMs can outperform generalist models in production environments where precision is non-negotiable. The Hybrid Future The reality isn’t LLM or SLM — it’s both. LLMs will serve as the reasoning engines, orchestrating complex workflows and bridging gaps across domains. SLMs will deliver deep expertise in critical tasks, ensuring enterprise-grade performance. This hybrid approach mirrors how organizations operate: broad leadership supported by domain experts. As AI adoption accelerates, companies that can strike the right balance between generalist adaptability and specialist efficiency will set the standard for the next wave of digital transformation. Question for you: In your industry, are you leaning more toward the power of generalist LLMs, the precision of SLMs, or a blended strategy?

Generative AI transforms how we work, learn, and solve problems. As auditors, we play a critical role in helping the organization balance innovation with risk. Here are some actionable steps to guide our organization, based on an article by Charles King in the Internal Auditor Magazine in December 2024: 1. Understand the Frameworks ↳ Familiarize yourself with AI risk management frameworks like NIST AI Risk Management Framework. ↳ This knowledge helps you assess AI risks effectively. 2. Establish Clear Policies ↳ Create an acceptable use policy for GenAI. ↳ Include what tools can be used, what data is allowed, and the usage limits. 3. Prioritize Training ↳ Employees need to understand both the benefits and risks of GenAI. ↳ Train them to use GenAI tools responsibly and to spot inaccuracies. 4. Use Secure Tools ↳ Enterprise-grade GenAI solutions offer better security. 5. Monitor and Adapt ↳ Implement systems to review GenAI usage regularly. ↳ Identify misuse, update policies, and adapt to new risks. What’s your favorite GenAI tool? Share in the comments. Reference: King, Charles. 2024. A Guide to GenAI. Internal Auditor Magazine December 2024 #internalaudit #ITaudit #digitaltransformation

Agentic AI is shifting from proof-of-concept to strategic capability. But for it to be enterprise-ready, it must evolve—fast. It’s no longer enough for agents to plan, reason, and act. Enterprises demand systems that are secure, efficient, observable, and accountable. Here’s what enterprise-ready Agentic AI really means: 🔒 Security by Design • Role-based execution and permission control • Tool-level sandboxing and isolation • Safe tool invocation and prompt injection prevention 💸 Cost Control and Efficiency • Lean prompt engineering and selective memory • Dynamic model routing (SLM > LLM when feasible) • Token-aware orchestration and call limits per agent/task 📈 Performance at Scale • Real-time responsiveness in agent chains • Lightweight planning loops with controlled recursion • Caching, precomputation, and optimized memory usage 🌱 Sustainability and Green Software • Emission-aware agent design (based on GSF SCI principles) • Green prompting and clean energy-aware execution • Monitor compute cost + optimize for energy efficiency ✅ Trust, Auditability, and Governance • Full observability of agent decisions and tool usage • Explainable reasoning paths and deterministic fallbacks • Human in the loop (when required) and compliance reporting (AI Act, SOC2, internal audits) 📊 Visibility and Observability • Dashboards for memory, latency, and token usage • Workflow heatmaps and traceable agent behavior • Integration with enterprise AIOps and monitoring systems ✅ Agentic AI must now be lean, secure, explainable, and scalable. It’s not about building more agents. It’s about creating the right ones—that last, that scale, and that earn trust. 🔍 For a deeper dive into designing cost-, carbon-, and complexity-efficient agentic systems, do visit https://leanagenticai.com #agenticai #leanagenticai

We are nearing the limits of the known antibiotic universe. For decades, progress has largely meant revisiting familiar molecules, even as resistance continues to outpace discovery. A recent effort from Massachusetts Institute of Technology changes the nature of the search itself. Instead of screening what already exists, researchers used generative AI to design tens of millions of hypothetical compounds that have never been synthesized or cataloged before. This is not deeper exploration of known space, but the creation of entirely new chemical territory. The AI generated molecules from first principles, guided by rules of efficacy and synthesizability. Several candidates that emerged are structurally unlike existing antibiotics and appear to act through a more fundamental mechanism: disrupting bacterial cell membranes. That distinction matters. Resistance often develops against drugs targeting specific internal proteins, but compromising the membrane is a broader, harder-to-defend strategy. In early studies, one AI-designed compound proved effective against drug-resistant gonorrhea by targeting a novel membrane-related protein, while another cleared MRSA infections in animal models, operating outside known antibiotic classes. The deeper shift here is conceptual. Generative models expand discovery beyond what can be searched or screened, into what can be designed. At a time when antimicrobial resistance is a growing global threat and the traditional pipeline is stagnant, this exploration-first approach offers a credible path forward. The next chapter of antibiotic development may depend less on rediscovery, and more on invention. #ArtificialIntelligence #GenerativeAI #DrugDiscovery #AntibioticResistance #AntimicrobialResistance #ComputationalBiology #AIinHealthcare #Biotech #LifeSciences #ScientificInnovation

OpenAI launches GPT-Rosalind to bring specialised AI reasoning into drug discovery: 🔘OpenAI has launched GPT-Rosalind, its first purpose-built AI model for life sciences, designed specifically for biology, drug discovery, and translational medicine rather than adapting a general model to scientific use 🔘The model is positioned as a reasoning engine for science, combining chemistry, genomics, and protein biology with the ability to navigate data, tools, and literature in a single workflow rather than treating each step in isolation 🔘In practical terms, it targets the hardest parts of early R&D such as understanding protein function, identifying drug targets, and predicting interactions, areas where failure rates are high and timelines can stretch to a decade or more 🔘Unlike typical AI tools, GPT-Rosalind is designed to actively support scientific workflows by generating hypotheses, retrieving evidence, and even suggesting experimental or chemical optimizations, effectively acting as a co-pilot for researchers 🔘Access is restricted to pharma, biotech, and research institutions, reflecting both the sensitivity of biological research and the need for expert validation in high-stakes domains like drug development 💬This signals a shift from general-purpose AI toward domain-specific reasoning systems in pharma R&D, where the competitive edge will come less from having AI and more from embedding it deeply into end-to-end scientific workflows #digitalhealth #ai #pharma

If you’re an AI engineer, product builder, or researcher- understanding how to specialize LLMs for domain-specific tasks is no longer optional. As foundation models grow more capable, the real differentiator will be: how well can you tailor them to your domain, use case, or user? Here’s a comprehensive breakdown of the 3-tiered landscape of Domain Specialization of LLMs. 1️⃣ External Augmentation (Black Box) No changes to the model weights, just enhancing what the model sees or does. → Domain Knowledge Augmentation Explicit: Feeding domain-rich documents (e.g. PDFs, policies, manuals) through RAG pipelines. Implicit: Allowing the LLM to infer domain norms from previous corpora without direct supervision. → Domain Tool Augmentation LLMs call tools: Use function calling or MCP to let LLMs fetch real-time domain data (e.g. stock prices, medical info). LLMs embodied in tools: Think of copilots embedded within design, coding, or analytics tools. Here, LLMs become a domain-native interface. 2️⃣ Prompt Crafting (Grey Box) We don’t change the model, but we engineer how we interact with it. → Discrete Prompting Zero-shot: The model generates without seeing examples. Few-shot: Handpicked examples are given inline. → Continuous Prompting Task-dependent: Prompts optimized per task (e.g. summarization vs. classification). Instance-dependent: Prompts tuned per input using techniques like Prefix-tuning or in-context gradient descent. 3️⃣ Model Fine-tuning (White Box) This is where the real domain injection happens, modifying weights. → Adapter-based Fine-tuning Neutral Adapters: Plug-in layers trained separately to inject new knowledge. Low-Rank Adapters (LoRA): Efficient parameter updates with minimal compute cost. Integrated Frameworks: Architectures that support multiple adapters across tasks and domains. → Task-oriented Fine-tuning Instruction-based: Datasets like FLAN or Self-Instruct used to tune the model for task following. Partial Knowledge Update: Selective weight updates focused on new domain knowledge without catastrophic forgetting. My two cents as someone building AI tools and advising enterprises: 🫰 Choosing the right specialization method isn’t just about performance, it’s about control, cost, and context. 🫰 If you’re in high-risk or regulated industries, white-box fine-tuning gives you interpretability and auditability. 🫰 If you’re shipping fast or dealing with changing data, black-box RAG and tool-augmentation might be more agile. 🫰 And if you’re stuck in between? Prompt engineering can give you 80% of the result with 20% of the effort. Save this for later if you’re designing domain-aware AI systems. Follow me (Aishwarya Srinivasan) for more AI insights!