Browsing: Guides & Tutorials

Most teams ask the wrong first question when they begin building LLM applications. They ask whether they should use LangChain, LlamaIndex, raw API calls, MCP, a custom agent framework, or whatever library is getting attention this month. That question matters, but it is not the architecture question. The better question is this: where should orchestration authority live? In enterprise AI systems, orchestration is not just chaining prompts together. It controls how the application routes work, retrieves context, calls tools, handles state, enforces policy, records telemetry, and recovers from failure. Once an AI application leaves the prototype stage, the orchestration layer…

Enterprise AI teams are learning a hard lesson: agents do not behave like human users. A human analyst may ask a few questions, inspect a dashboard, and refine a query. An AI agent may generate dozens or hundreds of intermediate queries, inspect schemas, test assumptions, call tools, create temporary state, retry failed paths, and coordinate with other agents. Once organizations move from chatbots to agentic workflows, the data layer becomes more than a place to retrieve documents. It becomes an operating substrate. That is why agentic data systems deserve attention. The next wave of enterprise AI will not be limited…

Scenario The VCF 9.1 upgrade path has been selected. The planning tool has been reviewed. The target version is understood. The team has a maintenance window on the calendar. That does not mean the environment is ready. A VCF 9.1 upgrade runbook needs to prove more than task order. It needs to prove that the organization can execute the upgrade, validate the result, stop safely when needed, and hand the platform back to operations without ambiguity. This is where many upgrade plans become too lightweight. They contain the sequence, but not the gates. They mention backups, but not restore confidence.…

Enterprise AI is moving from experimentation into production design. That shift changes the conversation. The question is no longer, “Can we run a model on-prem?” The better question is, “Can we operate private AI as a secure, governed, scalable platform inside the enterprise?” That distinction matters because private AI is not just about GPU access. It is about where data lives, how identity is enforced, how models are governed, how inference endpoints are exposed, how cost is controlled, and how operations teams support the platform after the proof of concept is over. For enterprises looking at on-prem private AI, three…

The practical controls that reduce context, model calls, loops, and unit cost while preserving task quality. GenAI prototypes rarely fail a cost review because of one expensive prompt. They fail when that prompt becomes a workflow. A production agent may add system instructions, conversation history, memory, retrieved documents, tool definitions, tool results, reasoning, retries, validation calls, and a final response. What appears to be one user request can become several model invocations before the task is complete. The visible prompt is only the beginning. The more useful unit of analysis is the entire agent run. That distinction matters because token…

Most enterprise AI conversations still start in the wrong place. Teams compare model benchmarks, argue over context windows, test one vendor against another, and treat the model as if it is the architecture. That might be enough for a demo. It is not enough for a production system that has to retrieve governed data, enforce access boundaries, call tools, produce evidence, log decisions, recover from failure, and survive audit scrutiny. The real shift is not only from smaller models to larger models. It is from model-centric AI to system-centric AI. A compound AI system uses more than one component to…

TL;DR A VCF 9.1 upgrade should not be treated like a normal maintenance window where each team updates its component, checks a dashboard, and leaves. The sequence matters. The ownership model matters. The readiness evidence matters. For VCF 9.1, VCF Operations is not a side task after the platform upgrade. Existing Aria Operations instances must be handled before core VCF components move forward, and VMware guidance requires Aria Operations to be at version 8.18 before it can be upgraded to VCF Operations 9.1. VCF Operations is also mandatory in VCF 9.x, which means environments without it need to account for…

Private AI is not just a question of where to place GPUs. That is where the conversation often begins, but it is not where the real architecture work lives. The harder problem is deciding how private AI will be operated, governed, secured, monitored, scaled, and recovered once it moves beyond the proof of concept stage. That is why VMware Cloud Foundation 9.1 deserves a serious place in the private AI discussion. VCF 9.1 is not interesting only because it can run AI workloads. It is interesting because it gives infrastructure teams a way to treat AI as part of the…

Human approval sounds simple until the agent is allowed to do something real. It is easy to put a button in front of an action and call it “human-in-the-loop.” Approve. Reject. Continue. Stop. That may be enough for a demo. It is not enough for an enterprise agent that can modify a ticket, send a customer message, restart a service, change a firewall rule, approve a refund, update a record, trigger a deployment, or delete data. The real design problem is not whether a human is technically involved. The real design problem is whether the approval path is scoped, timed,…

TL;DR The VCF 5.2.x to 9.1 upgrade is not just a component upgrade. It changes how teams should think about ownership. In VCF 5.2.x, many operational responsibilities were still organized around product lanes and appliances: SDDC Manager, Aria Operations, Aria Suite Lifecycle, Aria Automation, Log Insight, Identity Manager, vCenter, NSX, and ESXi. VCF 9.1 pushes the model toward a fleet-level management plane. That does not remove instance and workload-domain ownership. It changes the question from: “Who owns this appliance?” to: “Is this a fleet-level responsibility, an instance-level responsibility, or a workload-domain execution responsibility?” That distinction matters because VCF Operations, VCF…