Skip to content
BlendDesk

Pricing

Seats for your people. A meter for the AI, with a ceiling on it.

Two lines, one of which you can cap. Below is exactly what each one measures, what counts as a resolution, what un-counts it, and which fees are Meta's rather than ours.

There are no prices on this page yet.

We are still setting the figures, and we would rather publish nothing than publish a number we have to walk back. The model below is settled and it will not change when the figures land — this page simply gains a table. Until then we quote in writing, and we hold the quote.

Ask for a quote

The model

Two lines on the invoice.

Software is billed by the seat. AI is billed by the outcome, and only when the outcome happened. Nothing else on the platform is metered.

A seat is a person who answers customers
One monthly line per person who can open the inbox and reply. Seats are named, not concurrent, and you move a seat between people when someone leaves.
A resolution is a conversation Scout finished
The metered line counts resolutions, not messages, not conversations and not tokens. If Scout did not finish the job, there is nothing to count. The full definition is below, and it is the one we bill against.
Meta's message fees are not ours to bill
WhatsApp runs on your own Meta business account. Meta invoices that account directly for delivered messages. We never mark it up, because it never passes through us.

What we will never meter

  • Messages. Inbound, outbound, yours or your customer's.
  • Contacts in your workspace, however many you accumulate.
  • Conversations a person handled. A reply your team wrote is covered by that person's seat.
  • A credit balance that expires at the end of the month.
  • Storage of your own conversation history.

The definition

What counts as a resolution

Written out in full, because a metered AI line is only as trustworthy as the thing it meters. This is the definition we bill against.

One resolution is counted when, in a single conversation, Scout handled the request end to end and any one of these is true:

  • the customer confirms the request is handled;
  • the customer does not come back about it within 24 hours of Scout's last reply;
  • Scout completed a scenario or workflow that ends the request — the order was tracked, the slot was booked, the return was logged.

Counted once

Once per conversation, however many messages it took and however many tools Scout had to call. A customer who comes back about the same request inside 24 hours continues the same resolution; it is not a second one.

No resolution is counted when

  • Scout handed the conversation to a person, at any point, for any reason.
  • A guardrail blocked the action, or an approval was declined.
  • Scout told the customer it could not help, or asked them to wait for the team.
  • The run happened in the playground or inside an evaluation. Testing is never billed.
  • The conversation was spam, a duplicate, or a message with no request in it.

The 24-hour clawback

24h

If anyone on your team replies to that conversation within 24 hours of Scout's last message, the resolution comes off the bill. You do not file a claim and you do not argue it in a support ticket: it is removed automatically, and the run trace shows the human reply that removed it.

A human reply inside a day is the clearest signal there is that the AI did not finish the job. Billing for it would be billing for a failure.

Every counted resolution has a trace

What Scout read, which scenario matched, which tools it called, which guardrails it passed, what it drafted and who approved it. If you disagree with a counted resolution, the evidence is already in the record.

See a run trace

The ceiling

Set a monthly ceiling. We stop at it and tell you.

A meter without a ceiling is a liability, and the one thing a finance team cannot sign off is an invoice with no upper bound. So the ceiling is part of the product, not a support request.

A warning before, not an invoice after
You pick the ceiling and the point at which we warn you. Your admins get told when you cross it, inside the app and by email.
At the ceiling, Scout stops resolving
Conversations keep arriving and your team keeps answering exactly as they would with Scout switched off. Nothing is dropped and no customer is left on read.
Nothing above the ceiling is billed
Not as an overage, not as a credit, not on next month's invoice. The ceiling is the number, and we never silently overspend it.
Only you raise it
The ceiling changes when someone on your team changes it. We do not raise it for you, and we do not raise it to keep service running.
A spend-cap meter filling through the month: a warning at 84% of the ceiling, the ceiling reached, Scout stopping, conversations routing to the team, and a line confirming nothing above the ceiling was billed.

Illustrative demo with fictional data.

Meta's fees

Meta bills you for messages. We never mark that up.

WhatsApp connects through your own Meta business account and your own number, so Meta invoices you directly for delivered messages, priced by your customer's market and by template category. That invoice does not pass through BlendDesk, which means there is no percentage for us to add and no rate for us to inflate.

The 24-hour window

When a customer messages you, a service window opens for 24 hours. Inside it you reply with ordinary messages. Outside it you may only send an approved template, at template rates. Answering quickly is cheaper than answering late, which is one of the reasons the inbox is worth having.

The gap worth knowing about

In Egypt, Meta charges $0.0644 for a marketing template and $0.0036 for a utility template — about eighteen times. Ten thousand sends is $644 or $36 depending on which category the template was approved under. That is the largest single lever a merchant controls, and it is not a lever we or anyone else can pull for you.

Meta for Developers, USD rate card effective 1 October 2026, transcribed 2 November 2026.

Open the WhatsApp cost calculator

Getting a number

Tell us the shape of your month and we will put a price in writing.

No discovery call needed to get started — the planner below turns your answers into a note our team reads before the first conversation.

What we need from you

  • The channels you want connected, and whether your Meta assets already exist.
  • How many people will answer customers.
  • Roughly how many conversations you handle in a month.
  • Which jobs you would want Scout or an automation to take: order status, confirmations, cart recovery, qualification.

What you get back

  • A written price for seats and for the metered line, with the ceiling we recommend.
  • What is available today and what depends on Meta's approval, stated plainly.
  • The resolution definition on this page, unchanged. It is not negotiated per customer.

Sketch the setup you have in mind.

Pick the channels, team size and features you're curious about. We'll turn your choices into a note for our team, so the first conversation starts where you are.

Channels
Team size
Interested in

Your summary

Choose a few options and a summary of your setup will appear here.

This isn't a quote. It just helps us prepare for the conversation.

FAQ

Billing questions

Why are there no prices on this page?

Because the figures are not settled, and a price we have to retract is worse than a price we have not published. The model is settled: seats plus metered resolutions, with a ceiling. When the figures are ready this page gains a table and nothing else here changes.

What exactly is metered?

Resolutions, as defined above. Not messages, not contacts, not conversations, not tokens, and not anything a person on your team handled.

What if Scout resolves something badly and we have to step in?

If anyone on your team replies within 24 hours of Scout's last message, that resolution is removed from the bill automatically. The trace records the human reply that removed it.

What happens when we hit the spend cap?

Scout stops resolving and every conversation routes to your team, exactly as it would if Scout were switched off. Nothing above the ceiling is billed, and the ceiling only moves when you move it.

Do you take a cut of Meta's message fees?

No. Meta bills your own business account directly and we never mark it up. The calculator shows what Meta will charge, from Meta's published rate card.

Do we need Scout from day one?

No. The shared inbox works without it. Scout is a separate system that needs an AI provider configured for your installation, so you can add it when you are ready — and until you do, the metered line is zero.

Can we start with one channel?

Yes. Connect the inboxes you need now and add others later.

Do we need our own Meta accounts for WhatsApp?

Yes. WhatsApp connects through the Cloud API using your own Meta business account and phone number, which is also why Meta's fees are yours and not ours.

Let's talk about your setup.

Tell us about your team and the channels you use. We'll take it from there.