but they never miss an opportunity to miss. they could've been in a race with anthropic. can't believe they don't see it.
honestly not far off from the value demolition they did to poor heroku RIP. loved that product until they came in.
beyond my comprehension how badly salesforce/slack shit the bed with agentic ai.
slack should've been the data layer for everything agentic. it sits at the heart of the enterprise - all the data flows through it. it really is the data warehouse.
...
2/ Find a signal few people are using. The edge is in the data nobody else is watching yet.
3/ Differentiate the message(HARD), not the trigger. Same signal, sharper angle, real relevance.
Signal is a commodity now.
Outbound today is a race to bottom as everyone's using the same signals
My startup @withwillowai just came out of stealth. Within hours, dozens of tools were in my inbox trying to sell me something. All clearly triggered by the same "left stealth" signal.
...
That's the problem. When everyone's pulling the same signal in real time, it stops being MEANINGFUL. It's just NOISE.
Three ways to stand out:
1/ Wait. Let everyone burn through the signal, then show up when the inbox is quiet.
...
Last week: launched @withwillowai , the AI Basecamp for the enterprise.
Today: closing $7M seed, led by @HetzVentures.
Every infrastructure era had its access layer. AD. Okta. Agents need theirs now. That's Willow.
On the Economy Channel earlier this week, on enterprise AI adoption.
The unlock isn't another security tool. It's governance underneath.
Visibility for security. Control for IT. Adoption for the Enterprise.
Three teams. One control plane.
@withwillowai
Today weโre launching @withwillowai
A year ago we started building Webrix because we saw a gap nobody was talking about yet. AI agents were already getting access to production systems with zero governance. No audit trail. No kill switch. Security teams had no visibility into what these agents were actually doing.
Back then, most people looked at me like I was solving a problem that didn't exist yet.
Fast forward to today. Anthropic just shipped Managed Agents with credential isolation and sandbox separation baked in. Every major provider is now racing to add security to their agent platforms. The problem I was worried about a year ago isn't theoretical anymore. It's mainstream.
But here's what nobody is addressing yet. Enterprises don't run one provider. They have Claude agents, GPT agents, open source agents, internal tools. All hitting the same production systems. All governed separately. Or more likely, not governed at all.
Provider-level security is great. But it creates silos. Your security team doesn't need ten dashboards for ten providers. They need one governance layer across all of them.
That's what we've been building. And the problem has gotten so much bigger than what we originally set out to solve that the old name stopped fitting.
So today we're relaunching as Willow, a new site, and a much sharper focus on what enterprises actually need: unified visibility, access control, and revoke across every AI agent in production. Regardless of who built it or where it runs.
If you're a security or platform engineering leader dealing with this exact mess, I'd genuinely love to hear how you're handling it today.
Because from what I've seen talking to dozens of CISOs and SecOps leaders over the last few months, most people's honest answer is "we're not."
@GergelyOrosz Most ppl are YOLOing their way to prod rn.
If you don't want to burn the house down you must have a proper governance policy in place for any agent interaction with a system. This is a must and this is why we built @withwillowai
@teej_m This is why you cant let agents interact with apis directly. You need a control plane. You need guardrails. You need policies. Get an mcp gateway for the love of god!