I spent months cold calling local service businesses (trades, salons, small restaurants). Originally I was selling websites, but the conversations kept drifting into how they track customers.

What I heard over and over: - Their "CRM" is a notebook, a phone's recent calls list, or the owner's memory. - They've tried real CRMs. Most quit within weeks because setup felt like a second job. - What they actually want to know is simple: did more people call this month, where did they come from, and what should I do next to get more?

That pushed me to build something much lighter, mostly tracking calls and Google listing views and turning it into a short weekly to-do list, rather than a pipeline tool. (I'm the founder of that, not linking it here.)

Question for the CRM folks: for a one- to five-person business, what's the minimum version of "customer tracking" that actually sticks? Have you seen any setup an owner keeps using past month two?

We’re an independent insurance agency moving from InsuredMine to HubSpot and will be starting with one of the lower paid plans while we figure out our workflows with just 6 accounts for now.

Epic will remain our main system, but or now, we really just need HubSpot to help producers stay organized with prospects, pipeline, follow-ups, tasks/reminders and some basic marketing. We definitely don’t want to create another system that requires a ton of duplicate entry.

For anyone who has implemented HubSpot, what do you wish someone had told you before you started?

I'm interested in what people have learned actually using it. Things that caused headaches, setup decisions you regretted, features or limitations you didn’t realize would matter, or things you’re really glad you did early on.

We’re trying to keep the initial setup simple enough that people will actually use it, but flexible enough that we can build on it later.

Any real-world advice is appreciated!

Depois de 4 anos trabalhando ativamente na implementação de módulos e correção de falhas de segurança no Whaticket, desenvolvi minha própria versão do sistema atendendo as necessidades que encontrei no mercado ao longo dos anos.

Meu sistema está bem completo e bem estruturado. Tenho apenas um cliente (uma empresa com 15 funcionários) no sistema.

Gostaria de dicas para vender mensalidades do sistema. Sempre fui dev pra alguém mas nunca vendi meu próprio sistema. Tenho LP, tenho demo do sistema online mas não sei como encontrar o cliente ou como me posicionar para que o cliente me encontre. Vcs que vendem sistema, dicas pra como começar?

Ps: imagem meramente ilustrativa

tl;dr: Would building an open-source version of Clay's orchestration layer that uses your own models or LLM API keys be useful?

Hey all. I'm a data platform engineer, not a GTM person. But recently I was poking at Clay's free tier for a personal project. I wanted to enhance some data with Claygents but ran out of credits after a few dozen rows. Continuing was $167/mo, or $495/mo to automate it via webhooks, which felt like a big ask for what was mostly orchestrating model calls.

GTM data engineers at my previous company described a similar pain: they paid hundreds of thousands a month to orchestrate their enrichment waterfalls, even though the model calls went to their own Claude account, which they already paid for per call.

When I looked for alternatives, the hosted ones still priced on credits, and the open-source ones were either CLI tools you drive from Claude Code or templates built to showcase another platform.

I'm personally into building open-source alternatives to expensive SaaS for the same reason Tobi at Shopify talks about arming the rebels. Every cost in your stack limits a builder's ability to bootstrap their idea and is just another hurdle when competing against well-funded companies. This feels like a space where a quality open-source option could make a difference.

Before I dig deeper, I'd love to hear from people who do this every day. What would an open-source version need before you'd switch to it, and is there any reason it isn't worth building?

AI support vendors can test their core product, but every enterprise rollout introduces company-specific refund rules, policies, entitlements, identity checks, approval limits, integrations and escalation procedures.

For anyone deploying an action-taking customer support agent:

• Who owns acceptance testing: CX operations, QA, implementation or IT?
• How do you verify both the conversation and the resulting CRM, billing or account state?
• Are you using scripted UAT, sandbox accounts, shadow mode or limited production rollouts?
• What customer-specific failures only became visible after launch?
• Would a repeatable environment with synthetic customers, policies and connected systems materially help, or does your current process already cover this?

I’m trying to understand whether this is a painful recurring job or primarily a one-time implementation task. Any vendor recommendations currently solving this problem is also much appreciated!

Thank you!