Why We Use DenchClaw: Team Stories
The Dench team shares why they use DenchClaw themselves—real workflows, honest limitations, and what it's like to build and use your own product every day.
Why We Use DenchClaw: Team Stories
Every team building a product faces the same question eventually: do they actually use it themselves? For us at Dench, the answer is yes — DenchClaw is our primary tool for managing relationships, tracking the business, coordinating the team, and building on top of our own platform.
This is that story. Not a marketing piece — a real account of how we use DenchClaw at Dench, what works well, what we've had to work around, and what we've built that wouldn't have been possible with any other tool.
Kumar's Setup: Investor, User, and Operator Relations#
As the co-founder who handles most external relationships — investors, key users, press, potential hires — I run DenchClaw as my primary relationship operating system.
The objects I use daily:
Investors — YC partners, angels, VC firms we're in contact with. Every interaction is logged. Before any call, I ask the agent: "Summarize my last 3 touchpoints with [investor] and what we discussed." I've been in the middle of a fundraising process where I hadn't spoken to someone in 6 weeks, and the agent surfaced the context in 15 seconds. That preparation matters.
Power Users — The 50-ish users I talk to regularly about the product. Feedback, feature requests, use cases, referrals. Each one has an entry document with their specific workflow, what they've asked for, what we've shipped that addressed their needs. When we're prioritizing the roadmap, I query: "Show me the 10 feature requests that appear most frequently across power user documents."
YC Network — Alumni, batch founders, partner intros. The YC network is massive and relationship-dense. DenchClaw helps me maintain it without the cognitive overhead of trying to remember who knows whom, who I promised to intro, who asked for what.
The thing I couldn't do without: asking the agent "who in my network is in [city] I should catch up with before my trip" and getting a real answer. LinkedIn would take me 20 minutes to find the same thing, and the result wouldn't include the rich context I have locally.
The honest limitation: I can't share my DenchClaw workspace with anyone else (yet — team workspaces is coming). For the relationships I need a second person to see, I still have to export or brief them manually.
Mark's Setup: Customer Success and Growth#
I run our growth operations, which means I live in the pipeline side of DenchClaw more than anyone on the team. My setup is less philosophical and more operational.
The leads pipeline: Inbound from the Show HN post, from GitHub, from Twitter mentions — all normalized into a single leads object with clear stages. The most valuable view is "Hot Leads" — filtered to people who've starred the GitHub repo AND messaged us AND work at companies over 50 people. That intersection is the segment most likely to convert to paid Cloud users when we launch.
The retention pipeline: For users who've installed DenchClaw and are active, I track their engagement via our analytics integration and surface at-risk accounts in a daily Telegram briefing. Any user who hasn't logged in for 7 days gets flagged. Any user who asked a question in Discord that we didn't answer gets flagged.
What I've built on top of it: A custom Dench App that shows our week-over-week install numbers, active user count, GitHub stars trend, and Telegram bot user count — all on one screen. Data pulled from multiple sources, unified in one dashboard, no external analytics tool required.
The honest limitation: My biggest workflow pain is updating records from mobile. I use Telegram for quick notes, but if I want to update multiple fields at once, I still prefer the web UI. The mobile companion app (coming Q2) will fix this.
How We Use DenchClaw to Build DenchClaw#
The meta-pattern worth describing: we use DenchClaw to track user feedback that shapes DenchClaw's roadmap.
Every feature request anyone posts in Discord, files on GitHub, or tells us in a call gets logged as a feature_request entry in DenchClaw. Fields: what they asked for, who asked, how many people asked, current status (backlog/planned/shipped/declined).
When we do roadmap planning, the query is: "Show me feature requests with more than 5 requesters, status = backlog, sorted by requester count descending." The top of that list drives our sprint planning.
This creates a direct loop: community asks → we log it → it shapes the product → community gets features they requested → community trusts that feedback matters → community gives more feedback.
What We've Discovered Using Our Own Product#
Building and using DenchClaw simultaneously has taught us things we wouldn't have learned any other way:
Natural language queries take practice to get right. When we first started using DenchClaw, our queries were too vague. We learned to be more specific — not "show me leads" but "show me people with Status = Lead, added in the last 30 days, who haven't been contacted yet." We've shipped improvements to the agent's ability to handle ambiguous queries as a direct result of using it ourselves.
The daily Telegram briefing is the highest-value feature. Of everything in DenchClaw, the morning briefing has the most impact on our actual work. It's the thing that makes us feel like we're not missing anything. We've invested disproportionately in making it configurable and reliable as a result.
Customization is only valuable if you use it. The flexibility of DenchClaw — you can model anything — is also a trap. We've seen users get lost creating the perfect schema instead of using the tool. We've started shipping more opinionated starting templates as a result.
Data quality comes from frictionlessness, not process. We don't have a "CRM hygiene" meeting. We don't audit records. We have zero friction updates via Telegram, and our data quality is better than it was when we used HubSpot. The lesson: make updates effortless, don't try to enforce compliance.
The Tool We're Actually Proud Of#
We know DenchClaw has rough edges. There are features that still need polish, workflows that require more steps than they should, and edge cases we haven't addressed.
But when we use it every day and it surfaces a forgotten follow-up, finds a relevant contact for an intro, or builds a dashboard in 3 minutes that would have taken an afternoon in any other stack — we feel like we're building something genuinely useful.
The benchmark we use internally: "Would we use this even if we hadn't built it?" For our core workflows — investor relations, user research, growth pipeline — the answer is yes. That's the bar every feature has to clear before we ship it.
If you're curious about how we use any specific part of DenchClaw, come ask in the Discord (#founders channel). We're genuinely interested in how it compares to other people's workflows.
Frequently Asked Questions#
Do Dench team members use DenchClaw as their only CRM, or do you use other tools too?#
Primarily DenchClaw. We still use GitHub Issues for feature tracking (and sync them into DenchClaw), and we use email for external communication. But for relationship management — investor, user, partner, and network relationships — DenchClaw is the primary tool.
How do you handle handoffs when you're a small team and sometimes two people own the same relationship?#
We have a simple protocol: whoever is primary on a relationship owns the DenchClaw entry. If a second person has context, they log it as a note in the entry document. Because we can both read the entry document, there's no context loss when one of us is traveling or unavailable.
Will DenchClaw add native team collaboration features?#
Yes — team workspaces is on the Q3 2026 roadmap. Multiple users, shared database, attribution, @mentions. This is the most-requested feature from teams who've adopted DenchClaw.
Is using your own product as a team actually difficult when it has rough edges?#
Yes, sometimes. We hit bugs before users do (which is good for users). We occasionally lose patience with things that aren't polished. But the upside is that we prioritize the right things — the features that are rough are the ones we experience directly, so they get fixed faster. Dogfooding with real stakes is the best product development tool we have.
Ready to try DenchClaw? Install in one command: npx denchclaw. Full setup guide →