@hrdive

Over the past few weeks, I’ve seen troubling stats on “shadow AI” and spoken with leaders losing sleep over it. A recent @hrdive article reported that 45% of employees have used AI tools their companies banned and 58% have pasted sensitive data into them. The Fortune piece that.

What do you want to know?.

@glean 5. Don’t overindex on ROI. One of my favorite questions this week came from a Gleanie who asked me: “If the benefit exceeds the cost, can we do more hackathons?”. My answer was yes, even if it doesn’t. Most projects won’t ship, and that’s OK. The value comes from the few that do,.

@glean 4. No hacking on your OKRs. It's easy to fall back on what’s familiar: your goals, your roadmap, your comfort zone. We set a simple rule: anyone who hacks on their OKRs is disqualified. We want projects to push the status quo instead of protecting it.

@glean 3. Build for the long term. Hackathons are about moving fast. But if you only reward speed, you’ll probably get shortcuts. That’s why we have a “Codecraft” award to recognize durable, elegant code.

@glean 2. Don’t just chase moonshots. Big, bold ideas matter. But every company, including ours, has parts of work that have too much friction: manual handoffs, clunky workflows, outdated processes. That’s why we've created a Hackathon award for the project that best tackles the grungy.

@glean 1. Don’t limit them to engineers. Designers, PMs, data scientists, and GTM teammates all take part. The most innovative projects cut across functions.

Today is the final day of Hackathon Week at @Glean. I’m proud of the hard work and innovation our team has shown over the past few days. Hackathons are an important tradition at Glean. But they’re also an investment, and over time, we’ve learned a few things that make them most.

Two strikingly similar headlines surfaced this past week that should make every leader pause:. • “Companies Are Pouring Billions Into A.I. It Has Yet to Pay Off.” — New York Times. • “Companies Are Pouring Billions Into AI. Here’s Why They’re Not Seeing Returns” — Forbes. The

I’ve been reflecting on the GPT-5 launch. The pushback wasn’t just about performance. It revealed something bigger. Three key takeaways stood out to me: .1. The future of AI is multi-model. When models were pulled without warning or opt-outs, the frustration wasn’t only about.

@glean

Innovation doesn’t happen once a quarter – it happens every day. We ship improvements the moment they’re ready, so customers can start driving impact right away. Here’s what’s new in @Glean in this month’s drop:.𝗥𝗲𝗳𝗿𝗲𝘀𝗵𝗲𝗱 𝗔𝘀𝘀𝗶𝘀𝘁𝗮𝗻𝘁 𝗨𝗜: Conversational.

Support teams face constant pressure to resolve cases faster without overloading engineering. For one @Glean customer, valuable resources were tied up in avoidable tickets, MTTR (mean time to resolution) hovered at nearly two days, and agents spent hours manually triaging cases.

RT @glean: DAY 2 of our August Drop: Agent debug mode 🔍. Ever wonder how your AI agent comes up with answers? With debug mode, now you can….

RT @glean: DAY 1 of our August Drop: new Assistant and sidebar ✨. What's new?.✅ Quicker clarity through a refreshed UI.🤝 Powerful new brows….

GPT-5 is now live in @Glean, giving customers the chance to explore the model with the full context of their enterprise. Our AI eng team at Glean quickly ran an eval of @OpenAI GPT-5. Here’s what stood out: GPT-5’s broad tool use outperforms o3’s deep planning, achieving 83%

In enterprise environments, where agents must handle real workflows with real stakes – support tickets, policy updates, security incidents – quality isn’t optional. That’s why evaluation is a key part of the agent lifecycle. Standard LLM benchmarks aren’t enough. What matters is.

Link to article:

Meta’s vision for superintelligence – AI that augments human cognition – is powerful. But in the enterprise, individual superintelligence alone isn’t enough. Even if everyone operates at peak intelligence, the organization can still stall. Work is scattered across tools.

Wake-up call for every enterprise leader: If you don’t provide secure, enterprise-grade AI, your employees will turn to consumer tools – and that’s a risk you can’t afford. Here’s the reality: Shadow IT is real. Employees will use whatever tools help them get work done, even if.