60x

I sometimes get asked what makes Knowledge Graphs different to just using lots of MCP integrations. In the image below is a good example of the issues with plugging MCP connectors straight into the agent. Context fetching is not very efficient with MCP connectors. You retrieve a lot of context that isn't needed and at the same time also miss a lot of context which is needed with each call. This gets worse when lots of MCP connectors need to be used simultaneously for fetching context. This means that the output quality could be higher and the token spend could be lower. A Knowledge Graph enables this!

"I hope he hasn’t lost any of his novel ideas for the AI Brain." …ok, I was genuinely worried about Fergus McKenzie-Wilson when I heard he got injured. One thing about him: he's unusually good at coming up with novel ideas for solving big problems. For the past 6 months we've been building the "AI Brain" - a knowledge layer for any enterprise. It ingests information from a plethora of data sources and transforms it into a unified data layer for AI agents to easily traverse. From there we can build any sort of custom workflow one might desire! As Employee #1 and a founding engineer, I like that I can influence all aspects of the codebase. I can also make an impact on product, design, and infrastructure decisions that drive value for our first users. I'm glad Ferg is recovering fast. His ideas have shaped a lot of the AI Brain, and we're not done needing them yet!

9 out of 10 people wouldn’t spot this. This 60x founding engineer is so cracked that not everyone would even recognise what they're looking at. Before joining us, Alex Choi had already: • Launched 4 SaaS platforms • Launched 7 consumer apps • Won 6 hackathons He's a true end-to-end builder. Rare in a world where most people do one thing well. "No matter what we need Alex to do, he'll handle it." Literally. He's a gatekeeper against AI slop, a vibe-coding cleanup specialist with supercharged QA instincts. In just 6 months, he kept breaking his own record, Cursor-maxxing over $3,000/month, consistently. And how we met? Pretty romantic, actually. It started in Paris. We came together at the Louvre and won among 6,000 contestants at RAISE Summit, the world's largest AI hackathon. He built the WhatsApp agent. I built the orchestrator across all 6 agents. Roses are red. Violets are blue. 60x engineers meet through Unicorn Mafia. So should you.

I turned down a YC-backed startup to join a company founded by my hackathon teammates. The reason I chose 60x is simple: I back the founders. Charlie Cheesman, Fergus McKenzie-Wilson, and I won the RAISE Summit hackathon together the summer before. When they told me they had already bootstrapped to revenue, it was a no-brainer. Ferg is an exited founder. Charlie sold a business in only 20 hours. They have a track record of going from 0 to 1. And I believe in the vision: Enterprises have very different bottlenecks, and off-the-shelf products don't always solve them. Forward-deployed projects can often be the fastest way to unlock real value. We use an AI Brain as the base technology and can ship customised workflows in 60 hours. We’ve already won together, so I trust how we'll tackle whatever comes next. It's been a strong 6 months, and we're building with good vibes every single day! We're hiring. If you'd turn down a YC-backed offer for the right founders, you might be our kind of person.

Most enterprise AI fails because it lacks a map of your world. The simple solution: Knowledge Graphs. Right now, most AI systems use Retrieval-Augmented Generation (RAG) which relies on Vector Similarity. It finds text that looks like your question but doesn't understand the meaning behind it. Ask an agent to "audit all contracts for Project X." Vector search finds the name "Project X." But it fails to connect the Project to the people, emails, and history that actually drive it. Moving your internal AI from "chat" to "execution" requires structured reasoning. A Knowledge Graph maps your people and processes as a web of logic. Without a graph, your agent is a fast reader with no memory. In the last few months, 60X has built the knowledge core for some of the largest FTSE 100 companies. We turn siloed data (SharePoint, Slack messages, OneDrive) into an interconnected graph. TLDR: Architect the foundation that lets your software think

It’s been 6 weeks since Jaya Gupta surfaced the next big wave in AI. Context Graphs. AI systems do not fail because models are weak. They fail because they lack relevant context. Quality outputs only come from quality inputs. And most enterprise data is a complete mess. Data lives everywhere: In documents. In tools. In processes built over years. Most of all, it lives in decision traces: The exceptions, overrides, and ‘tribal knowledge’ hidden in Slack threads. The value is there. But it is not accessible to AI. This is why outputs feel generic. AI pilots stall, copilots disappoint. The next wave of AI impact will not come from bigger models. It will come from better context systems. This means building systems that understand: - Relationships between data - Meaning across documents - How entities connect over time We turn every exception into a searchable precedent for the next AI agent run. Context Graphs aren't just 'chains of thought.' They are durable, replayable records of how context turns into action. Incumbents see the data after the fact. We capture the full context at decision time, while it’s still actionable. This is how we move from AI pilots that 'disappoint' to autonomous systems that actually understand why a decision was made. 60x.ai builds the Context Graphs that make autonomous AI possible. The winners won’t be the companies with the most data. They’ll be the ones who made it meaningful first. ➕ Follow me, Fergus McKenzie-Wilson, to see what we’re building at 60x.ai.

"We need to implement AI." These words have killed more AI projects than I can count. Why? They put technology before outcome. They lead to great demos… and zero adoption. The best projects almost always start with a different question: "Where are we bleeding time, money, or customers?" From there, teams work their way backwards to the tech. Not the other way around. When you do this, here's what actually gets built: Automated reports: Every company has a report nobody wants to write. Automate it → now it takes minutes instead of days and you can iterate 5x more while maintaining quality. Churn detection: Churn signals are everywhere: product usage, emails, meetings. Impossible to catch manually at scale. Deploy analyst agents → get alerts before churn happens. Immediate revenue impact. Enterprise knowledge bases: Your team spends hours searching for info that sits across CRMs, meeting notes, and databases. Deploy a knowledge base → turn it into an internal search engine to get instant answers on deals, projects, anything. → Bonus: it becomes the foundation for every agentic system you build in the future. The highest-value AI projects rarely sound exciting. They sound boring. And that's exactly how you know they'll actually get used. What's the most tedious process in your organisation right now?

Stay tuned for more on enterprise Agent security!