# “I haven’t found a tool that does context as effectively as Unblocked does.” URL: https://getunblocked.com/customers/advidi/ Customer: Advidi Published: 2026-07-23T12:00:00.000Z Metrics: 400+ repositories reachable by people and agents through one MCP, Minutes to resolve decisions that used to stall in Slack for weeks, 1 MCP replacing a toolchain of token-hungry native connectors Advidi connected its engineers and its homegrown AI agents, like the OpsClaw incident bot, to one context engine spanning 400+ repositories, Slack, Confluence, and Jira, through a single MCP. --- Advidi is an affiliate marketing network in the ad tech space, and it's building its own affiliate tracking platform on top of a decade of history: some 300 legacy repositories, another hundred microservices on the new stack, and the knowledge that explains it all scattered across Confluence, Miro, repository READMEs, and Slack threads that get lost as time moves on. That knowledge gap hit two audiences at once. Advidi's engineers needed context to ship, and the AI agents the team was building needed it even more. CTO Jorden van Breemen's answer was to connect both to the same context engine. ## The challenge: smart people, smart agents, no context For people, every answer required another person. "It always involved somebody else," says Jorden van Breemen, Advidi's CTO. "You go to Slack and ask your peers, and maybe somebody links the right doc or points at the right repository. What it cost us, most of all, was time: your own, and everyone else's." Advidi wasn't just adopting AI tools; it was building its own agents. The flagship is OpsClaw, an operations agent built on OpenClaw that lives in Slack. When something breaks, the team loops OpsClaw into the thread the way they'd loop in a colleague. It crawls AWS, reads the logs, and investigates. But OpsClaw had a ceiling, and it was the same ceiling the humans had: it could see that a microservice was failing, but not the code or the decisions behind it. Cloning 300 repositories into an agent's context on every incident wasn't an option. The team tried the obvious fix for both audiences: wire up the native MCP servers for GitHub, Confluence, and Jira. It made the problem expensive instead of slow. "You get this extreme context bloat," Jorden says. "You load all of your Confluence documentation or crawl all of Jira. It was a token-expensive job, and you'd still reference outdated docs and miss things." Smart people, smart agents, and the context both needed locked away in twenty places. That's the problem Jorden flew to San Francisco to solve. He went to every context talk at the conference; one of them connected directly to the tools Advidi uses and the problems it has. "I gave it a shot," he says, "and never looked back." ## The solution: one context engine for people and agents > "Unblocked combines all of the tools we use into one product, one single MCP. And the search is efficient: the context it gathers is relevant and recent. I haven't found another tool that does that as effectively as Unblocked does." > Jorden van Breemen, CTO, Advidi _[Video: One MCP, permissions, efficient search]_ Every alternative Jorden evaluated did one of those things. None did all three. And the permission layer is what made it safe to hand the same engine to people and agents alike, because Advidi's context lives in Slack, and often it never leaves. "If a conversation happens in a channel where only a subset of people are present, somebody else cannot loop that context into another conversation," Jorden says. "Maintaining that privacy while still exposing people to everything they need is unique, and very valuable to us as a company." **For people:** Unblocked runs as the MCP inside Claude and ChatGPT, is looped into CLAUDE.md files so coding assistants pull institutional knowledge automatically, and answers directly in the Unblocked app and in Slack when something breaks and nobody knows which of 400 repositories to open. **For agents:** Jorden connected OpsClaw to Unblocked and scoped its access just as he would for a human teammate. Unblocked gave the agent visibility into the GitHub, Confluence, Jira, and Slack knowledge relevant to operations, and nothing more. OpsClaw received the same institutional knowledge as the rest of the team, governed by the same permission model. ## OpsClaw, unlocked: from log line to code line With Unblocked connected, OpsClaw stopped being an agent that could only describe symptoms. > "I'm not going to clone 300 repositories and let a bot crawl through them every time an incident happens. Now an incident occurs, it reads the logs, sees the 500 response, and it can point at exactly the right place in the codebase. It lets people tackle problems they simply couldn't before." > > — Jorden van Breemen, CTO, Advidi _[Video: OpsClaw pinpoints incidents in code]_ It changed how incidents work socially, too. Before, an investigation meant engineers looping in their own AI assistants or reading code alone. Now the whole team watches OpsClaw work in one Slack thread and looks at the answer together. The agent isn't a private tool; it's a colleague in the channel. OpsClaw isn't alone. Advidi runs OpenClaw analytics bots connected to Looker and BigQuery, where users can ask questions across hundreds to thousands of tables. "This is most of all a context problem," Jorden says. The bots know the schema; Unblocked supplies the meaning: the Confluence page that defines what a metric actually measures, the Slack conversation that's relevant right now. And the pattern extends to people who never touch an IDE. Product owners refine tickets with indirect access to the code: they were given GitHub access not to read code, but so Unblocked could read it for them, showing what a user story touches across the chain and turning Slack refinement threads into development-ready stories. Business users ask about anything from architecture to the HR handbook. When a technical debate stalls in Slack, the team tags Unblocked for a third, sometimes tie-breaking opinion that has actually looked at the code, resolving in minutes questions that used to be postponed for days, weeks, or indefinitely. ## The results: a new way of working with context Netting it out, Jorden doesn't describe a better search box. He describes a structural change: the team stopped maintaining glossaries and index maps in every repository just so people could find where to look. Context stopped being a scavenger hunt, for everyone and everything that needs it. > "It saves tokens on scanning repositories, and it saves a lot of time for the people who use it. Before, they had to scour every place context lives in our company. Now Unblocked guides them in the right direction: in development, in data, in product refinement, and in the business problems people have." > Jorden van Breemen, CTO, Advidi _[Video: A new way of working with context]_ Everyone at Advidi who works seriously with AI now uses Unblocked daily, Jorden included; he calls himself its most frequent user in the company. Asked what would happen if it disappeared tomorrow, his answer is the shortest in the interview: "I'd have to look for something else that handles context the way it does, and I haven't found it yet. We'd see a dip in velocity and go back to the problems we had before." ## The bottom line for engineering leaders Advidi didn't change how it works to get here, and it didn't write a single new page of documentation. It connected the tools it already had (GitHub, Confluence, Jira, Slack, Google Drive) to one context engine, and let everything downstream draw from it safely: engineers, product owners, business users, and the agents the team builds next. > "Everyone has context living in multiple places: code, Slack, conversations. It's always hard to pull it back into one. Unblocked solved that for us: ask a question, get all the relevant context from your tools, with permissions maintained. It's a problem a lot of companies struggle with today." > Jorden van Breemen, CTO, Advidi Get in touch with us today to see how Unblocked can be the context engine for your engineers and your agents.