19 stories
Executive summary AI agents open up new routes to enterprise data, and app-level permissions can't cover all of them. In this article, Cisco engineering leader Greg Laws [https://www.linkedin.com/pulse/policy-based-access-agent-era-greg-laws-zaozc/] walks through how policy-based access keeps data…
Every AI pipeline runs on small decisions. Is this ticket about billing or shipping? Does this prompt need the expensive model? Should this agent be allowed to delete that file? Most teams hand each of those calls to a frontier model like ChatGPT. At 10 requests a day, nobody notices.
Most AI workflows run on a single model. Every step goes to the same large language model, whether the task is drafting a customer reply or answering a simple question like "Is this ticket urgent?" That setup gets expensive fast.
Most of the AI calls inside your software don't need an essay. They need an answer: which team owns this ticket, whether a tool call is safe, whether a lead is worth chasing. The usual approach is to ask a large language model for a paragraph and then parse it into something your code can use.
Board week has a familiar rhythm. Finance pulls numbers from three different systems, someone rebuilds the variance table by hand, and the cashflow forecast is already a day old by the time it lands in the deck. The pressure is real.
A regulator, a customer, or your own security team asks one question after an AI system gives a wrong answer: what exactly led to it? If your logs only show a prompt and a response, you have nothing to hand over.
Ask two people to build the same dashboard and you'll likely get two different definitions of "churn," three follow-up meetings, and a Slack thread that dies long before the numbers do. That's the real cost of solo dashboard building.
Ask an AI agent to pull a number, then ask a follow-up question about that same number, and watch what happens. In a lot of enterprise setups, the agent forgets what it just told you.
Introduction I onboarded in a haze of wiki tabs, stale Notion pages, and Slack DMs that went unanswered for hours. You have seen it too. The new person who wants to ship but spends three weeks hunting down a policy doc or the right MongoDB connection string.
Introduction Last quarter I watched a colleague toggle between Slack, Snowflake, and Salesforce for forty minutes to answer a single question from a customer success manager. The data was all there. The context was not.
Introduction You know the drill. You are deep in a problem, headphones on, finally making progress, when Slack pings. "Hey, what's the SOC 2 shared responsibility doc again?" Twenty minutes later, you have lost your flow, answered three more questions, and forgotten what you were building.
Seven people can collaborate in one thread, every question asked, every decision made, and every dead end explored immediately visible to the next person who needs it. Nobody has to ask "where's the file," because the context was never locked inside a single, private chat history.
Most AI tools were built for one person at a time. You open a private chat, teach it something, and close the tab, and everything you just taught it stays with you. That model made sense when AI was a personal productivity boost.
You've been here before. The same alert fires again, the fourth time this quarter, and you're the only one who knows exactly what to run to fix it, because you're the one who figured it out during a panic-driven 3 AM call months ago.
A finance team once caught an agent approving a 40% discount [https://atlan.com/know/context-graph-vs-knowledge-graph/] for a client capped at 25%. The model sounded confident. It cited the right policy name. It was dead wrong.
A team runs an AI agent on what looks like a clean dataset and asks a simple question: what's the projected Q4 revenue for the East Coast division? The agent returns a number, confident and completely wrong. The real culprit usually isn't the model.
You've felt it. You tell Claude your product roadmap, then jump into ChatGPT to draft a launch email and realize it has no idea what you're talking about. The knowledge you painstakingly built in one tool is completely absent in another.
Your team lives in Slack. Your answers live in a warehouse, a CRM, and a dozen other SaaS tools that don't talk to each other. So people toggle tabs, copy-paste numbers into a channel, and hope the figure they just posted is still current by the time someone reads it.
Showing 18 of 19 stories
Load more storiesGet the top AI news and tools straight to your inbox every week.
Weekly · AI news from PromptQL Blog
Join 15,000+ AI enthusiasts
