Understanding Context in Artificial Intelligence

𝗗𝗲𝘀𝗶𝗴𝗻𝗶𝗻𝗴 𝗖𝗼𝗻𝘁𝗲𝘅𝘁-𝗔𝘄𝗮𝗿𝗲 𝗔𝗜 𝗔𝗴𝗲𝗻𝘁𝘀: 𝗧𝗵𝗲 𝟲 𝗗𝗶𝗺𝗲𝗻𝘀𝗶𝗼𝗻𝘀 𝗼𝗳 𝗖𝗼𝗻𝘁𝗲𝘅𝘁 Building AI agents isn’t just about fine-tuning prompts or plugging in APIs. The real differentiator lies in how effectively we design and manage context. Context defines the agent’s role, behavior, reasoning, and decision-making. Without it, even the best models act inconsistently. With it, agents become reliable, explainable, and enterprise-ready. Here are the 6 essential types of context for AI agents:  1. 𝗜𝗻𝘀𝘁𝗿𝘂𝗰𝘁𝗶𝗼𝗻𝘀 – Define the who, why, and how: • Role (persona, e.g., PM, coding assistant, researcher) • Objective (business value, outcomes, success criteria) • Requirements (steps, constraints, formats, conventions)  𝟮.𝗘𝘅𝗮𝗺𝗽𝗹𝗲𝘀 – Demonstrate desired (and undesired) patterns: • Behavior examples (step sequences, workflows) • Response examples (positive/negative outputs)  𝟯.𝗞𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 – Embed domain and system understanding: • External context (business model, strategy, systems) • Task context (workflows, procedures, structured data)  𝟰.𝗠𝗲𝗺𝗼𝗿𝘆 – Extend reasoning across time: • Short-term memory (chat history, state, reasoning steps) • Long-term memory (facts, episodic experiences, procedural instructions) 𝟱.𝗧𝗼𝗼𝗹𝘀 – Extend capability beyond training data: • Tool descriptions act as micro-prompts • Parameters and examples guide usage 𝟲.𝗧𝗼𝗼𝗹 𝗥𝗲𝘀𝘂𝗹𝘁𝘀 – Close the loop by feeding outputs back into reasoning: • Orchestration layers attach results • Enables agents to adapt dynamically 𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀: By designing across all six dimensions, we move beyond “prompt engineering” into structured context engineering. This makes agents: • More autonomous • More explainable • Easier to scale across enterprise systems In practice, this framework underpins everything from agent orchestration protocols (MCP, A2A) to multi-agent architectures in production. Question for you: When building AI agents, which of these six contexts have you found most challenging to implement at scale?

Context engineering is quickly becoming one of the most critical skills in applied AI. Not prompt tweaking. Not model fine-tuning. But knowing what information a model needs, and when to give it. That is the real unlock behind AI agents that actually work. At its core, context engineering is about delivering the right information to the model, at the right time, in the right format; so it can reason effectively. It pushes developers to think more intentionally about how they shape a model’s inputs: 🔸What does the model need to know for this task? 🔸Where should that information come from? 🔸How do we fit it within the limits of the context window? 🔸And how do we prevent irrelevant or conflicting signals from getting in the way? Why does this matter so much? In practice, most agent failures are not due to weak models. They happen because the model did not have the context it needed. It missed a key fact, relied on stale data, or was overloaded with noise. Context engineering addresses this directly. It forces you to design the flow of information step by step, not just what the model sees, but how and when it sees it. This context can come from many places: 🔹Long- and short-term memory (such as prior conversations or user history) 🔹Retrieved data from APIs, vector stores, or internal systems 🔹Tool definitions and their recent outputs 🔹Structured formats or schemas that define how information is used 🔹Global state shared across multi-step workflows Frameworks like LlamaIndex, LangGraph AI, LangChain, are evolving to support this shift, giving developers the tools to manage context with much more precision. And there are now better resources than ever to help teams write, select, compress, and organize context with real control. Image from Langchain blog. #contextengineering #llms #generativeai #artificialintelligence #technology 

Context engineering is becoming one of the most important skillsets in the AI era — because the quality of an AI system’s output depends entirely on the quality of context it receives. This framework breaks down the six pillars that shape how AI understands, reasons, retrieves, and responds with accuracy and relevance. Here’s what each component contributes: 🔹 Prompt Techniques - Tree of Thoughts (ToT): Helps the model explore multiple reasoning paths and choose the optimal answer. - ReAct Prompting: Blends reasoning (“think”) with action (“act”), allowing the model to use tools, gather data, and refine responses iteratively. 🔹 Memory - Short-Term Memory: The model’s immediate context window — everything currently “in view” that shapes its next step. - Long-Term Memory (ReAct-based): External vector storage that allows AI to remember past interactions, facts, and patterns over time. 🔹 Retrieval Retrieval pipelines break queries into chunks, embed them, fetch relevant knowledge from vector stores, and feed enriched context back into the LLM for more accurate generation. 🔹 Query Augmentation LLMs rewrite vague or incomplete queries into precise, structured prompts — enabling better problem-solving and more accurate output. 🔹 Agents AI agents reason step-by-step, use tools adaptively, access memory, decompose problems, and switch strategies dynamically when one approach fails. 🔹 Tools External tools expand an AI system’s capabilities — enabling database queries, API calls, file operations, search, and multi-step workflows through structured integrations. Context engineering is the hidden layer that transforms AI from a simple text generator into a reliable reasoning engine. When these six components work together — prompting, memory, retrieval, augmentation, agents, and tools — AI systems become dramatically smarter, more accurate, and far more capable of solving real-world problems. ♻️ Repost this to help your network get started ➕ Follow Prem N. for more

For years now, prompt engineering shaped how people worked with large language models. It was about finding the right phrasing to get predictable outputs. That approach worked for small tasks, but as models turned into agents that plan, use tools, and retain memory, the limits became obvious. One of Anthropic’s latest articles “𝘌𝘧𝘧𝘦𝘤𝘵𝘪𝘷𝘦 𝘤𝘰𝘯𝘵𝘦𝘹𝘵 𝘦𝘯𝘨𝘪𝘯𝘦𝘦𝘳𝘪𝘯𝘨 𝘧𝘰𝘳 𝘈𝘐 𝘢𝘨𝘦𝘯𝘵𝘴”, introduces the next phase in this evolution, called context engineering. It explains that success now depends on how well we manage what goes inside the model’s attention window rather than how we word instructions. Anthropic describes context as everything the model sees while reasoning, including prompts, data, retrieved results, tool outputs, and message history. Every token consumes a portion of the model’s attention, and as the window expands, its focus gradually weakens. The new challenge is to curate that space carefully. Below are the main lessons from Anthropic’s work that stand out for anyone building practical AI systems. 1. Treat context as a limited resource. Adding more information does not improve accuracy. Use only what directly supports the current reasoning step. 2. Write system prompts like structured briefs. Divide them into clear parts for background, instructions, tools, and expected output. 3. Build small, distinct tools. Each tool should solve one problem and return compact, unambiguous results. 4. Use a few canonical examples instead of long lists of edge cases. Examples should teach reasoning, not overwhelm the model with detail. 5. Retrieve data just in time rather than all at once. Lightweight references such as file paths or queries keep the model’s focus clear. 6. Compact long interactions. Summarize the conversation and restart with the essentials so that the model stays coherent over long sessions. 7. Store information outside the context window. Structured notes or state files help maintain continuity across projects. 8. Use sub-agents for large tasks. Specialized agents can work on details while a coordinator manages direction and synthesis. 9. Balance autonomy with reliability. Some data should stay fixed for consistency, while other parts can be fetched dynamically when needed. 10. Focus attention on signal, not volume. Every token should contribute to the next action or decision. Prompt writing will still matter, but the real skill now lies in shaping context and deciding what enters the model, what stays out, and how information evolves as the agent works. The next generation of LLM Agents will depend less on clever wording and more on precise design of memory, retrieval, and context. Context engineering is becoming the foundation for reliable agents that think and act across long horizons with consistency and purpose.

🔥 The Bundesliga Breakthrough: Personalizing Context, Not Content At the Sports Forum at Amazon Web Services (AWS) re:Invent 2025, I learned something rare… something most business leaders never hear: Everyone talks about personalized content. Almost no one talks about personalized context. And that’s where the Bundesliga — One of the top football leagues — is quietly years ahead of the market. Here’s the insight most executives miss: AI doesn’t scale content. AI scales understanding. Bundesliga’s AI system works because it doesn’t start with the match. It starts with the fan. Before a single line of commentary is generated, the system builds a real-time “context graph”: - who the fan is - how long they’ve followed the league - what commentary style keeps them hooked - which players they track - what cultural cues resonate in their region - and what emotional tone they respond to This is the whole magic. Gen AI is just the surface. The real breakthrough is the context engine underneath. >> Why this matters for executives? Most companies try to personalize by tweaking the message. Bundesliga personalizes by changing the lens through which the customer sees the message. That shift is massive. Because when you personalize context: - Engagement stops being random - Marketing stops being guesswork - CX stops being generic - Loyalty stops being an accident >> The uncomfortable question Most leaders ask: “How do we create more personalized content?” The better question — the Bundesliga question — is: “Do we truly understand the context our customers live in?” Because here’s the uncomfortable truth: > You can’t personalize content at scale unless you personalize context first. Bundesliga shows the future. The next decade of CX belongs to companies that invest not only in storytelling… …but in systems that understand their customers better than customers understand themselves. Your turn: 👉 How could your customer experience improve if your systems learned their context the way Bundesliga’s does? #CustomerExperience #AILeadership #GenerativeAI #Bundesliga #DigitalTransformation #Personalization

🧩 The Context Engineering Problem in AI One of the biggest problems in AI right now is context. Let me break this down in terms of what we're building at Pascal AI Labs Our early customers are public and private equity investment funds. Let's say this fund went and hired the smartest person in the world. Even with that high IQ, they’d still need to understand how the fund operates - its investment philosophy, research processes, and documentation norms before they could meaningfully contribute. So this new high IQ hire would spend months grounding themselves in the fund's institutional memory: how sectors behave, how decisions are documented, how insights are shared. For a human, that could take 3, 6, 12 months - sometimes years. The mistake many teams make with AI is assuming that a generalised LLM can skip that process. Most teams we encounter buy an enterprise license, plug it in, and expect it to think like their team. Inevitably, they get stuck at simple use cases like summarizing transcripts or retrieving snippets, and then hit the valley of disappointment. The hype quickly fades because the model doesn’t understand their world. The only real way to solve this problem is to give AI context — to make it part of your fund, not just a co-pilot. That starts with two steps: 1️⃣ Teach it the domain. Horizontal models are still unreliable on financial accuracy. We’ve seen customers try using them for deep research and end up with results that are only about 70% correct — and the problem is, you never know which 70% because the answers all _sound_ smart. So first, the system needs a foundation of financial context: how industries behave, which metrics matter, where to find them, what commentary is relevant, and so on. 2️⃣ Give it your institutional memory. Just like a first-year analyst, the AI needs access to everything that defines how your fund operates — internal models, memos, meeting notes, research documents, all of it. Without that, it can’t mirror your reasoning or outputs. At Pascal AI, we work on both steps. Our system runs on top of the best horizontal models and adds the scaffolding required to understand financial context. Once we connect a fund’s internal data, the system can analyze and interpret how that fund truly operates through our proprietary knowledge graph — the institutional backbone that maps how your fund actually works. Pascal AI makes AI a first-class citizen of your fund by adding the context required for it to operate at the same level as an analyst. So when you ask the system to analyze a company, it doesn’t just look at public data. It recalls your historical notes, trades, past commentary, and how you’ve thought about that sector before. It understands your investing style and generates insights within your unique context, not from a blank slate. It's very likely that the next wave of AI won’t replace analysts. It’ll work like one - shaped by your data, your memory, and your context.

Mid-conversation with Claude yesterday, I got this message: "Compacting our conversation so we can keep chatting. This takes about 1-2 minutes." At 62% capacity, I watched it reorganize its thoughts. And I realized: most AI users have no idea this is happening. Here's what you need to understand. Context windows are AI's working memory. It's the total text the model can "see" at once—your prompts, its responses, uploaded documents, everything. Claude offers 200,000 tokens (roughly 150,000 words) for paid users. Sounds massive until you're deep into a complex project. When you hit that ceiling, something has to give. Claude's approach: Auto-compaction kicks in around 95% capacity. Earlier messages get summarized, keeping what the AI thinks matters most. Your full history is preserved—you can scroll back—but the AI's "working memory" gets compressed. Each compression cycle loses granularity. Manus AI takes a different path. Rather than compacting in place, it externalizes memory to the file system—creating todo.md files to maintain focus, saving intermediate results externally, spinning up sub-agents with their own context windows for discrete tasks. When context fills up, it uses "recoverable compression"—dropping content but keeping URLs and file paths so it can retrieve information later if needed. Neither is perfect. Both involve tradeoffs. The takeaway: Context limits are real constraints on how much complexity AI can handle in a single session. If your team uses AI for research, strategy, or extended projects, you need to understand this. Three practical tips: → Checkpoint manually at 70% rather than waiting for auto-compaction at 95%. You control what's preserved. (This only works in Claude Code using the /compact switch) → Summarize at natural breakpoints. Ask the AI to capture key decisions before moving on. You may manually bring this across to another chat or ask it to save it to memory. → For complex projects, externalize documentation (E.g use Projects in Claude and ChatGPT). Don't rely solely on conversation memory. As context windows expand—Claude's testing 1 million tokens for some API users—this will matter less. But for now, understanding your AI's memory limits is the difference between productive collaboration and frustrating repetition.

Your LLM can write emails. But can it reason over your Q3 pipeline? The difference is context. I've been observing something interesting in the enterprise AI landscape: "All of the value in the market is going to go to CHIPS and what we call the ONTOLOGY." Most people focus on the chips part. They miss the ontology revolution. What am I saying? Follow me here. Ontology isn't just a philosophical concept. It's a data relationship retrieval mechanism to augment reasoning that structures, connects, and contextualizes it so AI can actually make sense of it. Think about it this way: Models are only as smart as the world they understand. Most LLM systems are trained on generic web-scale data. They're incredibly intelligent. And completely clueless about your business. The breakthrough comes when you ground AI in domain-specific ontology. When you transform business knowledge into structured, machine-understandable intelligence. This is the missing architecture layer. What ontology actually looks like: - A curated, evolving graph of business concepts. - Accounts, pipeline stages, personas, intent signals, forecast dimensions. - Not just data points. Relationships and meanings. Why it matters technically: It acts as a semantic engine enabling deeper reasoning, causality, and traceability. Not just prediction, but explainability and control. Raw signals become interpretable actions. The architecture stack: Raw data inputs flow into the ontology layer. The ontology layer feeds structured context to reasoning systems. Reasoning systems power coordinated agent actions. Most enterprises are missing this middle layer. They're connecting raw data directly to AI models. No wonder their agents make decisions that ignore core business logic. The companies getting this right understand something fundamental: - Entity Relationships: AI that knows how deals, reps, products, and timelines actually connect. - Business Rules Integration: AI that respects ownership hierarchies, escalation paths, and approval flows. - Action-Agent Mapping: AI that understands which specialist should handle which situation. This isn't about making AI smarter. It's about making AI business-aware. When agents operate from unified business context: - Decisions become coordinated across systems - Silos disappear between teams - Accuracy increases while blind spots reduce The result? AI that doesn't just automate tasks. AI that understands the business it's operating within. At Aviso AI, we've embedded this ontology layer at the core of our agentic architecture. It's what enables our agents to reason, act, and collaborate across GTM ecosystems. Because if compute is the fuel, ontology is the map. Context isn't just king. When it's structured as business ontology, it's transformational.

Ontology, Context, and the Foundation of Enterprise Intelligence Ontology defines the what, where, and why of data. It gives meaning, structure, and direction. While knowledge graphs once carried that banner in what I call “traditional/academic AI,” ontology has become the bridge that connects human understanding with enterprise intelligence. We’re now entering a phase where ontology meets reality. As more teams learn how to use RAG models, vector embeddings, multimodal systems, and natural language pipelines, we’re seeing a new wave of hybridization. Terms like Graph RAG, Graph ML, Graph Analytics, and the merging of AI Ops and ML Ops are everywhere. The naming is fragmented, the infrastructure is fragmented, and many organizations are still trying to cobble these pieces together. The truth is, without a unified foundation, nothing scales. You can’t build intelligent systems on inconsistent context. Ontology creates that foundation. It connects systems, people, and purpose through shared meaning. Knowledge graphs once promised this but often failed in practice. Three to six months after being built, they became outdated. People left, projects shifted, language evolved, and those scatterplots of meaning lost their relevance. Ontology, however, adapts. It represents not just static connections but living context that evolves with the organization. The future of enterprise AI depends on contextually aware intelligence. That means understanding how your company thinks, operates, and differentiates. It means defining what your enterprise knows, how it learns, and why it matters. ▫️When you know your what, where, and why, your systems gain clarity. ▫️When your foundation aligns with your ontology, you gain scalability. ▫️When you unify meaning, you build intelligence that lasts. #ontology #RAG #humanfirst #enterprise Forbes Technology Council Gartner Peer Experiences InsightJam.com PEX Network Theia Institute VOCAL Council IgniteGTM IA FORUM SSON Intelligent Automation Community Solutions Review 𝗡𝗼𝘁𝗶𝗰𝗲: The views within any of my posts, or newsletters are not those of my employer or the employers of any contributing experts. 𝗟𝗶𝗸𝗲 👍 this? Feel free to reshare, repost, and join the conversation!

As financial institutions accelerate their adoption of AI, one pattern has become increasingly clear: we are not constrained by model capability. We are constrained by our ability to give those models the right context to operate in complex, regulated environments. Most large enterprises now run dozens of models across credit risk, fraud, marketing, operations, and customer service. Yet very few can reliably provide an AI system with the foundational elements required for high-stakes decision-making: - A clear, unified representation of the customer or entity - An accurate understanding of what is happening in real time - The relevant product, risk, and regulatory constraints that define what actions are permissible Without this context, even highly capable models remain brittle and inconsistent. They may perform well in isolation, but they struggle in the dynamic workflows that define financial services. The challenge is not intelligence—it’s situational awareness. I believe this aligns closely with Ilya Sutskever’s recent observation that the era of performance gains driven purely by scaling is coming to an end. Scaling has produced exceptionally powerful general-purpose models, but it has not solved the problem of enterprise-specific reasoning. The next breakthroughs will come from new architectures and methods that allow models to use context more effectively, not simply from increasing parameter counts. To make AI reliable and responsible at scale, financial institutions must focus on building what I refer to as a context fabric: - a consistent way to represent customers, accounts, relationships, and events; - a structured approach to encoding policies, constraints, and guardrails; - and standardized task schemas that define exactly how AI systems should operate across workflows. This shift—from model-centric to context-centric AI—is essential for achieving the resilience, explainability, and trust demanded in our industry. It is not optional. It is the foundation for AI systems that can be deployed safely and deliver measurable business value. The real competitive advantage in the next phase of AI will belong to institutions that master context: not just the next model, but the infrastructure, governance, and reasoning layers that make AI truly enterprise-ready.