The per-seat pricing model was logical when every user needed a login and a dashboard. You paid for seats because humans were the operators.
That model is starting to break down. And it's creating an arbitrage opportunity for teams moving fast.
The Original Logic: Seats as Operators
In 1995, a CRM seat cost what it cost because that seat represented a human operator. A salesperson logging in daily, reading pipeline data, updating notes, making calls, advancing deals. The software was a tool they used directly. More salespeople = more seats = more cost.
The pricing structure made sense. You paid for what you used — human time and attention. The vendor invested in servers, support, features, and infrastructure for each additional user. More users = more load = more cost.
This logic held for 25 years. It still makes sense for most teams.
But something has changed: AI agents now handle 50-80% of CRM interactions at many companies.
The New Reality: AI as Operators
Think about what happens in a modern workflow:
A customer sends an email. An AI agent reads it, understands the context, pulls the relevant deal from the CRM, enriches the contact with recent activity, drafts a response, and flags the email for your salesperson to review.
A deal closes. An AI agent writes a summary, updates the deal stage, sends congratulations to the team, creates the next steps, and logs activities across related contacts.
A prospect goes silent for 30 days. An AI agent detects the pattern, drafts a follow-up, schedules it, and marks the deal at-risk.
Data entry, pipeline updates, contact enrichment, meeting summaries, follow-up drafts, pattern detection, exception alerts — in many companies, agents now handle the majority of these operations. The salesperson reviews and supervises. The agent operates.
The Math That Breaks the Model
Let's say you have a 10-person sales team. In the traditional model, you need 10 seats.
But here's what's actually happening: you need 10 human seats, plus 3-5 agent seats running in the background.
With per-seat pricing, you're paying for all 13-15 seats.
Let's put numbers on it:
Salesforce Sales Cloud (Enterprise): $165-$330/user/month
10-person team: $1,650-$3,300/month = $19,800-$39,600/year minimum
HubSpot Professional: $100/user/month
10-person team: $1,000/month = $12,000/year
Open Tooling CRM: $0 software + ~$20/month VPS
10-person team: $240/year
The arbitrage isn't close. It's a 50-150x difference.
Now, you might say: "But Salesforce and HubSpot include things Open Tooling doesn't." True. Let's be honest about that.
What You're Actually Paying For
HubSpot Professional and Salesforce Enterprise include:
Uptime guarantees (99.9% SLA): You don't have to maintain infrastructure. If something breaks, it's their problem.
Professional support: You call them, they fix it. You don't have to hire an ops person.
Compliance certifications: SOC 2, GDPR, HIPAA (on higher tiers). If you're regulated, this matters. They've already passed the audit.
Ecosystem integrations: Thousands of pre-built plugins, Zapier, Slack, native mobile apps, etc. Plug and play.
Feature richness: Email automation, marketing campaigns, customer service tools. It's all built in.
These are real value-adds. They have a cost.
But here's the question: do you need all of them?
Who Needs to Switch (And When)
The case for moving to self-hosted, open-source CRM gets stronger as your company grows or your use case diverges.
Startups burning runway on SaaS bills. Every dollar counts. If you have a technical founder or CTO, you can absolutely trade support SLAs for cost. You're self-hosted, you're responsible, you save $12,000-$40,000 per year. That's real runway extension.
Agencies managing multiple client CRMs. If you're running 5 separate CRM instances for 5 clients, HubSpot multiplies your cost 5x. Open Tooling lets you deploy multiple instances cheaply, even with different customizations per client.
Companies where AI agents are already doing the work. If half your "users" are agents, and you're paying $100-150 per agent per month, you're being arbitraged. Agents don't need seats. They need API access. That's free on Open Tooling.
Teams that need data ownership. If you're in healthcare, finance, or any regulated industry, having your data on your server (or your cloud account) might not be optional. HubSpot's cloud-hosted model doesn't fit. Open Tooling self-hosted does.
Companies building custom workflows. If your business is construction, recruiting, field services, or anything outside the "standard pipeline" model, you're fighting HubSpot's defaults. A headless CRM and a custom interface might be faster and cheaper than customizing HubSpot.
None of these are edge cases anymore. They're real patterns.
The Longer Prediction: Per-Seat Pricing Will Evolve
I think per-seat SaaS pricing has 3-5 years left as the dominant model in enterprise software.
Here's why: as AI agents become standard operators, per-seat pricing becomes absurd. You can't charge $150/month for a seat that's an automated task runner.
The pricing models that will replace it:
Usage-based: You pay for API calls, data stored, compute cycles, etc. Slack is starting to move here with seat+usage. This makes sense when agents are operators — you pay for what they do, not how many of them exist.
Outcome-based: You pay based on results (deals closed, revenue generated, etc.). This is harder to implement but it's how the best agencies already work.
Infrastructure cost passthrough: You pay for the server, the data storage, the bandwidth. Nothing else. Open Tooling is already here.
Companies moving to self-hosted, open-source infrastructure now are ahead of that curve. When the industry catches up, they'll already have the data ownership, the cost structure, and the architectural flexibility that future models require.
The Honest Tradeoffs
I'm not saying every company should leave HubSpot or Salesforce. I'm saying the economics are changing, and some teams are being over-charged by the old model.
If you move to Open Tooling, here's what you're trading:
You get: zero software cost, unlimited API access, full data ownership, agent-first architecture, no seat taxes, data hosted on your server
You lose: vendor support SLAs, compliance certifications (you provide your own), email automation (you integrate it), mobile apps (you build them), out-of-the-box marketing features
For teams with technical depth, that's a good trade. For teams without it, it's not.
The Real Arbitrage
The real opportunity is this: in 2-3 years, when it becomes obvious that per-seat pricing doesn't work for AI-first operations, the smart teams will already be self-hosted.
They'll have spent 6 months setting up Open Tooling, getting comfortable with the API, building custom interfaces, integrating it into their stack.
And then when the next platform launches with usage-based pricing (because that's the only way to stay competitive), they'll already have the infrastructure and the thinking to migrate easily.
The teams still on Salesforce or HubSpot will be locked in, accustomed to the per-seat model, and suddenly struggling to justify the cost.
Or they'll just pay it, because they always have.
Why We Built It This Way
We built Open Tooling CRM self-hosted and open-source specifically because we think this shift is coming.
We could have built a SaaS with a free tier, charged per seat at scale, followed the playbook. It would be easier to sell.
But that would have been optimizing for the past, not the future. The future is agents as primary operators, self-hosted infrastructure, and data you own.
Try it yourself. Deploy it locally in 10 minutes. See if the API makes sense for your team. If it does, you're ahead of the curve.
If it doesn't, HubSpot and Salesforce are still great products. They're just optimized for a different era.
