I am currently running an RFP. HubSpot, Salesforce, and 2 agencies are all on the table. Proposals are coming in, demos are scheduled, contracts are being drafted.

And somewhere in the middle of all this, a question keeps interrupting me.

If I am willing to learn how to build workflows in n8n and KNIME, and I already use Claude Code to write the logic, do I actually need a CRM at all? Or do I need what a CRM does — and I have just assumed, for years, that the only way to get it was to buy one?

This is not a conclusion. I do not have a working alternative running yet. This is the list of questions I am asking myself before I sign anything, and I think more digital leaders should be asking them too.

1. Am I Buying a Platform, or Buying Three or Four Specific Capabilities?

A CRM bundles a lot: contact records, pipeline stages, email sequencing, reporting, integrations, permissions, support tickets, sometimes marketing automation on top.

I rarely use all of it. Most teams do not. So the real question is not "which CRM is best." It is: which 3 or 4 capabilities am I actually paying for, and could each one be built as a standalone workflow?

Worldline needs transaction visibility. Winarco needs issuing and back-office sync. Schenk needs terminal-level reconciliation. None of that is "CRM" in the classic sense — it is data moving between systems on a schedule, with rules. That is exactly what n8n and KNIME are built for.

2. What Am I Actually Paying For — the Software, or the Fact That I Cannot Build It Myself?

This is the uncomfortable one. HubSpot and Salesforce pricing is not really pricing for software. It is pricing for the assumption that building and maintaining the workflow yourself is harder than paying someone else to have already built it.

That assumption used to be true for me. It is less true every month I work with Claude Code. If I can describe a workflow in plain language and get working automation logic back, the barrier to "build it myself" just dropped — and the SaaS premium I am paying did not.

3. What Happens When the Vendor Changes Their Roadmap, Not Me?

Every long-term CRM user has a story about a feature that got deprecated, a pricing tier that got restructured, or an integration that quietly broke after an update nobody asked for.

When I build the workflow myself in n8n or KNIME, the roadmap is mine. Nothing changes unless I change it. That is not a small thing for a company running cross-acceptance between Tamoil and partner stations, coordinating four separate stakeholders, where stability matters more than features I am not using anyway.

4. Can I Actually Maintain This Long-Term, or Am I Trading a Subscription for a Liability?

This is the question that keeps me honest. A CRM subscription is annoying, but it is predictable. A self-built workflow stack is only worth it if I can actually maintain it — not just build it once and hope.

So before I rule anything out, I am asking myself: do I have the time, and the team, to own this the way a vendor would? If the honest answer is no, that changes the calculation completely, no matter how good the proof of concept looks.

5. Is This About Saving Money, or About Owning the Logic?

If this were only about cost, the math might not even favor building it myself once I account for my own time. But that is not really what is pulling me here.

It is ownership. A workflow I build in n8n, with logic I designed and can read, is something I understand completely. A CRM is a black box I am renting access to. For a project this complex — Worldline, Winarco, Schenk, Management, all moving at once — understanding the system beats trusting the system.

6. What Is the Actual Switching Cost If I Am Wrong?

Last question, and the one that protects me from getting too excited about my own idea. If I build this myself and it does not hold up, what does it cost to walk back to a vendor CRM? Data export, retraining, lost time — all real.

If that cost is low, building it myself is a low-risk experiment. If that cost is high, it is a bet I need to be much more certain about before I make it.

Where This Leaves Me

I am not telling you to cancel your CRM. I am telling you I am asking myself harder questions before I renew mine — and the RFP responses I am reading right now are being judged against a self-built alternative, not just against each other.

That alone has changed how I read every proposal on my desk.

If you are running a similar evaluation, or you have already gone down the build-it-yourself path with n8n, KNIME, or Claude Code, I would like to compare notes. Connect with me or send me a message — I want to hear what you found before I make my decision.

Note: These reflections are my own personal thinking, not a position or recommendation from Tamoil. The RFP referenced here is being run on its own merits, independent of this post.