Hello.
Kicking off the Gauss blog — what we'll write about, and what to expect.
We're building Gauss because we spent years hand-building AI agents at previous companies — a marketing agent here, a GTM agent there, an engineering assistant somewhere else — and every one of them ended up needing the same substrate underneath.
Every team we've worked with has variations of the same three problems:
- The AI teammates you can buy off the shelf are prepackaged personas. They know what a "customer support agent" is in the abstract. They do not know what a customer support agent for your team is. Your escalation rules, your naming conventions, the three edge cases your CS lead cares about — none of that ships in the box.
- The internal alternative is a full ML team. For most teams that's not a real option. The ones who have it end up shipping bespoke Slack bots that break on the first refactor.
- The middle ground — a low-code workflow tool — doesn't think in terms of agents. It thinks in terms of triggers and steps. Fine for routing a Typeform to a Slack channel; wrong shape for "read this thread, decide what to do, take four different tool calls to do it."
Gauss is the substrate for the middle ground. Playbooks in plain markdown, universal tool access via MCP, scheduled routines, full audit chain — shaped so anyone on the team can build, not just the engineer who happened to draw the short straw.
What we'll write about here
- Field notes — what teams are actually doing with Gauss. Playbook patterns that work, ones that don't.
- Design decisions — the reasoning behind product choices. When we said no to per-seat pricing, when we said yes to MCP, why we picked Slack as the first surface even though it's not the last.
- Comparisons — where we sit relative to n8n, Glean, Slack AI, custom GPTs, and the "just give the intern a Zapier account" alternative.
- The occasional deep dive — how prompt caching actually works, what the audit chain protects against, what BYOK enables.
Nothing here will be corporate. If we get something wrong, we'll say so. If we change our minds on a design decision, we'll write about why.
Subscribe by hitting the RSS feed at /blog/rss.xml, or check back — we'll ship these when there's something worth saying, not on a schedule.
— The Gauss team