- Flat pricing (AI assistants) Definition
- a subscription where one monthly fee covers full product use without metering individual AI actions. Example: the inbox side of alfred_ (triage, drafts, follow-ups, the morning brief), which doesn't draw on credits. Contrast with credit-based plans where each automated step consumes quota.
- Usage pricing (AI assistants) Definition
- a model where cost scales with consumption, task credits, AI tokens, automated runs, or metered minutes. Common in agent builders and AI scheduling tools. Monthly bills become hard to forecast because the same subscription behaves differently in a light week vs. a busy week.
You wouldn’t accept a human assistant who invoices extra every time you forward an email. Yet that’s how much of the AI assistant market prices today: a monthly fee plus a meter that runs faster the more you delegate.
This piece explains why that structure fails professionals, where flat pricing wins, and when usage pricing is actually reasonable.
The Credit Trap in One Week
Picture a normal Tuesday that turns heavy:
- 80 inbound emails (launch week, client escalation, scheduling ping-pong)
- An agent or workflow triages and drafts on each meaningful thread
- Calendar AI reschedules three blocks to fit a new deadline
On a credit plan, that Tuesday isn’t “using the product as designed.” It’s a burn event. Users describe the same pattern across Lindy and Motion:
- Sign up because the headline price looks reasonable
- Automate real workflows in week two
- Hit credit anxiety, hesitate to turn on another agent or re-run a failed workflow
- Get a bill that doesn’t match the mental model of “$X per month”
Lindy Plus at $29.99 per user per month includes 3,000 credits, and Lindy lists everyday asks at 2 to 250 credits each. At 50 emails/day, a reactive triage agent can exhaust a month’s allowance fast, not because you misused the product, but because your inbox exists.
Motion Pro AI advertises $19/month but ships 7,500 AI credits per seat for scheduling actions. Heavy auto-reschedule weeks consume quota you didn’t price in when you compared “$19 vs $25.”
The product punishes the outcome you bought it for: less thinking about admin work.
Three Ways Usage Pricing Hurts Consumers
1. Unpredictable bills break trust
SaaS works when finance and founders can forecast spend. Credits turn a subscription into a variable utility bill, closer to AWS than to the email client it replaces.
You cannot answer “what will this cost me in Q4?” until you’ve run production volume for a full month. See Lindy pricing hidden costs: credit burn stays invisible until real triggers fire.
2. You ration the assistant during peak load
Peak load is when an executive assistant earns its keep. Usage pricing inverts the incentive:
| Week type | What you need | What credits do |
|---|---|---|
| Light | Maybe ignore the tool | Cheap, feels fine |
| Heavy | Maximum automation | Expensive, cap risk |
Users report not re-running failed agents, disabling steps, or avoiding experiments, behavior that only shows up on metered plans.
3. “Cheap” entry prices are comparison tricks
A $19/month sticker with 2,000 credits competes against a $29.99 plan whose inbox work isn’t metered in spreadsheets. It shouldn’t, not until you map credits to your email volume and workflow depth.
Best Lindy alternatives exist largely because searchers want simpler and cheaper, i.e. predictable. That’s a market telling you the pricing model failed the job-to-be-done.
Meter the Automation, Not the Assistant
alfred_ splits it on purpose. On every plan, the always-on inbox work doesn’t use credits:
- Content-aware email triage
- Voice-matched draft replies (you approve before send)
- Follow-up tracking and task extraction from threads
- Chat, in the app or by text
- Morning brief across email and tasks
What does use credits is the optional automation you choose to run: routines (ready-made jobs, like planning your week) and workflows (jobs you build yourself). Each plan comes with a monthly amount, from 2,000 credits on Starter to 25,000 on Max, and they refresh on your billing date. If they run out, routines and workflows pause until they refresh; chat, triage, drafts and the morning brief keep working.
Why we split it this way: the work that follows your inbox volume is the work a meter punishes most, so it stays off the meter. Routines and workflows are consumption you control, closer to the “optional and sporadic” case below, so that is where credits apply. The amounts are on the pricing page and the rules are on How credits work.
You should want alfred_ to touch every message that matters, not budget how many touches you can afford. That is the test for anything an assistant meters.
Comparison: Flat vs Credit Models (2026)
| alfred_ (credits on routines and workflows only) | Lindy (credits) | Motion (credits + seat) | |
|---|---|---|---|
| Headline price | From $29.99/mo | $49.99/mo Plus (credit count unpublished) | $19/mo Pro AI (7,500 credits/seat) |
| Bill predictability | Fixed plan price, credits refresh monthly | Depends on workflow steps | Depends on reschedule volume |
| Busy-week behavior | Inbox work costs no credits; routines and workflows draw on your monthly amount | Credits deplete faster | Credits deplete faster |
| When automation stops | Routines and workflows pause until credits refresh; chat, triage, drafts and the brief keep going | Credits exhausted | Credit cap / upgrade |
| Setup | Connect email + calendar | Build & debug agents | Tasks + calendar rules |
| Best for | Inbox + calendar overload | Custom multi-app agents | Auto task scheduling |
This isn’t “alfred_ wins every category.” Motion is strong for aggressive task-to-calendar scheduling. Lindy is an AI teammate for teams that work in Slack. The pricing comparison is about what you’re buying daily, if it’s email and calendar execution, meters work against you.
When Usage Pricing Is Fair (We’re Not Pretending It Never Works)
Usage pricing makes sense when:
- Consumption is optional. You run jobs sometimes, not always-on on your inbox
- Marginal cost is transparent, developers paying per API token know what they invoked
- The buyer controls volume, batch uploads, not “how many emails clients send you”
It’s a poor default when the pitch is “your AI executive assistant”, always-on, reactive, tied to communication volume you don’t control.
Per-seat pricing ($29/user/month) is fine. It’s predictable. The consumer-hostile pattern is seat fee + opaque credits + automation halt without a clear mapping from credits to outcomes.
How to Evaluate Pricing Before You Subscribe
Use this checklist (extends our AI assistant decision framework):
- Map one real week, count emails that would hit triage/draft, agent runs, or calendar AI actions.
- Ask the vendor, what happens at 100% of credits? Hard stop, overage, or upsell?
- Price the busy month, not the trial week, trials are artificially light.
- Compute $/hour saved, divide monthly cost by hours recovered; ROI beats sticker price.
- Prefer flat if the tool is infrastructure, if skipping a day means dropped balls, you need unlimited-normal-use pricing.
What “Flat” Should Mean (Red Flags)
Not every “Pro plan” is truly flat. Watch for:
- Credits rebranded as “AI actions”
- Separate AI subscription on top of base CRM/calendar fee
- Feature gates, drafts OK, triage premium
- Overage invoices not disclosed on the pricing page
alfred_ meters on purpose and says where: only routines and workflows use credits, the monthly amounts are on the pricing page, and what happens when they run out is written on the credits page.
The Bottom Line
Usage pricing for always-on work assistants transfers your volume risk, inbox spikes, launch weeks, crisis threads, back onto you. It trains rationing at the wrong time and makes “AI assistant” feel like a utility you meter instead of a role you delegate.
The test is product design honesty: if we say we’ll handle your email and calendar, you shouldn’t pay more when we succeed. So the always-on inbox work isn’t metered, and the optional automation on top is, with the amounts stated up front.
