HubSpot

What it is

HubSpot is the default SMB and mid-market CRM. Sales Hub, Marketing Hub, Service Hub, and Data Hub share one record graph — companies, contacts, deals, tickets, and custom objects in a single schema. Breeze is the AI brand layered over that data: Assistant inside the app, plus agents that HubSpot is consolidating into Agent Hub. Most ops teams still get more out of pairing HubSpot with vertical AI tools (Clay for enrichment, Gong for conversations, Claude and n8n for orchestration) than out of Breeze alone.

Note the 2025 renames if you are reading older docs: Operations Hub is now Data Hub, Lists are Segments, and Copilot is Assistant. The product is the same; the vocabulary in every pre-2026 runbook is not.

Why it shows up in every RevOps stack

• One record graph. Deals, tickets, and custom objects sit in the same schema with bi-directional sync to most adjacent tools — no middleware to keep marketing and support looking at the same account.
• Two official MCP servers. The remote CRM server at mcp.hubspot.com went GA on 2026-04-13, with write access to contacts, companies, deals, tickets, line items, products, and engagements, and read access to campaigns and pages. A separate local Developer MCP server (GA 2026-02-19) drives app and CMS builds from Claude Code, Cursor, or VS Code. Very few CRMs in this segment ship vendor-maintained MCP at all.
• Workflows plus custom code actions. Native automation covers the bulk of RevOps logic; Node.js custom code actions take the long tail — but see the Data Hub line below, because that tier is where they live.

Pricing reality

Sales Hub list, checked 2026-09-04:

• Free — $0, up to 2 users
• Starter — $20/seat/mo month-to-month, $7/seat/mo on annual commit; 500 HubSpot Credits
• Professional — $100/seat/mo monthly, $90/seat/mo annual; 3,000 credits; $1,500 one-time onboarding, required
• Enterprise — $150/seat/mo annual; 5,000 credits; $3,500 one-time onboarding, required

Two structural details do more damage to budgets than the headline seat price. First, the seat type: those rates buy Sales Seats. Teammates who only need CRM access take Core Seats at $50/mo (Professional) or $75/mo (Enterprise), and view-only seats are free — a 12-person GTM org with 6 quota carriers is not 12 × $90. Second, HubSpot Credits are the AI meter. They cost $0.010 each, or $9 per 1,000 bought annually, they reset monthly with no rollover, and the allowance is not additive across products.

Data Hub is a separate line and the one teams forget: Professional is $720/mo on annual commit ($800 monthly), includes 1 Core Seat and 5,000 credits, and caps webhook triggers at 100; Enterprise is $2,000/mo with unlimited webhook triggers and 10,000 credits. Programmable automation is gated here, not in Sales Hub.

Real-world band: annual multi-hub commits typically land ~20–35% under list, an estimate from reseller and pricing-aggregator triangulation rather than a published rate. Onboarding fees are also negotiable, usually through a Solutions Partner. Budget Sales Pro plus Data Hub Pro as the realistic mid-market floor.

Best for

RevOps leaders at $1M–50M ARR B2B SaaS standardizing sales, marketing, and CS on one CRM, who want strong native UX and are willing to trade Salesforce-grade configurability for a system their team will actually keep clean.

Not for you if your primary requirement is programmable automation at a low seat count — you clear $720/mo for Data Hub Professional before writing a single custom code action, and a $99/mo n8n instance against the HubSpot API does the same job. Also not for you if you need custom objects with deep validation logic and territory-level permission models; that is a Salesforce build.

Versus the alternatives

• Salesforce — pick it when configurability and enterprise governance outrank time-to-value: custom objects with real validation rules, complex territory hierarchies, and an admin who owns the org full-time. It costs more in implementation than in licenses.
• Pipedrive — pick it when you need a sales pipeline and nothing else. Roughly a third of HubSpot’s per-seat cost at the working tier, with none of the marketing or service surface. Downgrading later is not realistic, so choose it only if you are confident marketing automation stays out of scope.
• Attio — the fastest-growing AI-native entrant in this segment. Pick it when your data model is unusual (communities, PLG accounts, investor pipelines) and you would otherwise fight HubSpot’s opinionated schema. Its integration surface is a fraction of HubSpot’s, so it is the wrong pick if the CRM has to be the hub for six other systems.

Watch-outs

• Credits are the meter, and agents eat them. Breeze agents draw on the same monthly allowance that resets with no rollover; monitoring-style agents run on the order of ~100 credits per monitored contact per month (estimate), so a 3,000-credit Professional allowance covers roughly 30 prospects. Guard: run any agent on one segment for a full billing cycle and read the credit report before enabling it portal-wide, and buy top-ups annually at $9/1,000 rather than on demand at $0.010.
• Programmable automation is a second purchase. Custom code actions and webhook triggers live in Data Hub Professional, not Sales Hub. Guard: price the Data Hub line into year one, or deliberately keep the long tail in n8n or Zapier against the CRM API and never buy it.
• Seat-type drift. Teams provision Sales Seats by default because that is what the rep who asked for access had. Guard: audit seat types against actual usage each renewal — the Core and view-only downgrade is the single largest recoverable line on most HubSpot invoices.
• Reporting ceiling around $10M ARR. Native reporting stops answering cohort and attribution questions. Guard: plan the BI plug-in (Looker, Hex, Mode) as part of the CRM budget rather than as a surprise, and keep source-of-truth definitions in the warehouse.
• Breeze trails purpose-built tools. Enrichment, conversation intelligence, and analysis are all a generation behind Clay, Gong, and Claude. Guard: treat Breeze as convenience automation on data you already have, and keep the specialist tools for anything a revenue number depends on.