Seats are the floor.Usage is the ceiling.
AI features, automations, exports — every new release ships with a meter behind it. Your seat plan was never built to carry them. Seats, usage, and credits on one ledger. One invoice. Margin pinned to every workspace.
Four things modern SaaS actually needs.
Join seats and meters per workspace. Compose hybrid plans without a deploy. Gate AI features at the SDK. Re-cost the book when a provider price changes. One ledger, wired in once.
Cost, by workspace. Seats and usage in one row.
The $99/seat plan looks identical at signup. Six months in, one workspace is running 4M AI completions and 80GB of exports. The other is logging in twice a week. Joining seats and meters per workspace tells you which one is paying for the other — before renewal.
- Seats × AI calls × storage × exports per workspace. One ledger, not four CSV exports.
- Flag underwater accounts before renewal. By plan, by feature, by AI model.
- Re-rate any month against any future plan. "What if AI completions had a fair-use cap?"
- Notebook-CoBusiness · 64 seats$1,920+272%
- Brightside HRPro · 22 seats$660+161%
- HatchworksPro · 18 seats$540−190-240%
- LoomstackBusiness · 40 seats$1,200−2814%
- Forge MonoEnterprise · 320$14,400+366%
- Gridline OpsBusiness · 50 seats$1,500−48-22%
- seat$29.00/ seat / mo
- meter[ai]$0.012/ call · free_tier: 500 × seats
- meter[storage]$0.18/ GB / mo · free_tier: 25GB
- credit[trial]1,000applied · workspace_age < 30d
- billingmonthlyin_arrears · usage_capped
Mix seats, usage, and credits in one plan.
Modern SaaS plans bill three motions at once: seats for access, meters for AI and storage, credits for trial and overage. Describe the plan in English; ship it versioned, without a deploy. Pricing stops being a quarterly engineering project.
- Compose any plan shape. Seat, request, GB, credit — combine per plan.
- Versioned at the workspace level. Grandfathering is a flag, not a migration.
- A/B prices, then keep what wins. Promote the version that lands the margin you want.
Gate every AI feature at the SDK. Not the dashboard.
The AI summarizer is in the Pro tier. The bulk export is metered. The new agent is rolling out to 5% of workspaces. Cap or override on the hot path — billed when they overflow, lifted with a flag instead of a deploy.
- Entitlement checks on the hot path. Atomic check-and-deduct on every request.
- Block, or auto-bill the burst. Overflow is a line item, not an outage.
- Per-workspace override from the dashboard. Lift a cap for one customer. No deploy, no PR.
- ✓Notebook-Co· ai.summarize(doc)$0.02allowed
- ✓Brightside HR· export.csv(rows=12k)$0.18allowed
- ✗Hatchworks· ai.agent.run(steps=40)$1.40capped└─ pro tier · monthly AI call cap reached
- ✗Loomstack· export.bulk(rows=4M)$3.20capped└─ business tier · 1M row monthly cap
- ✓Forge Mono· ai.embed(rows=80k)$12.40override└─ enterprise · burst until 17:00 · billed @ $0.0001/row
- Freetier · 01gpt-4o-mini→haiku-3.5+28%margin
- Protier · 02gpt-4o→sonnet-4+14%quality
- Businesstier · 03sonnet-4→sonnet-4—no change
- Enterprisetier · 04opus-4→opus-4🔒contract
Swap providers re-pricing your plans.without
OpenAI drops a new model. Anthropic cuts price 40%. Your storage moves from us-east-1 to a cheaper zone. Update the rate in your Cost Sheet and every new event is re-costed — the price on your plans doesn't move.
- Per-provider cost rates. Different tiers, different providers, same ledger.
- Plans are versioned and reversible. Roll back v9 of your provider map without a migration.
- Forecast margin before you commit. Margin trend and forecast per workspace.
Stop guessing.
Start knowing.
Wire Nozle into your stack and see your first margin report in 60 minutes.
- Setup
- 60 min
- Card required
- no
- Sales call
- no
- First invoice
- in 60 min