Headless 360 pricing, budgeting before the price list exists
Salesforce has posted no standalone Headless 360 production price, so the real cost runs through Agentforce Flex Credits and per user add ons, and the budget is built from the meters underneath. This guide separates the anchors you can trust from the four lines still undisclosed, and the one move that beat them all: a capped unit rate, locked for the term.
Prepared by Redress Compliance · August 6, 2026 · Salesforce advisory. Based on the Agentforce and Data Cloud order forms reviewed 2025 to 2026.
Executive summary
The anchors are real, the price list is not.
As of mid 2026 there is no standalone Headless 360 production price, and none is likely: the cost flows through existing meters, and the trustworthy anchors are $500 per 100,000 Flex Credits, 20 credits per Agentforce action, and the $125 per user per month add on.
Everything billed will decompose into those numbers or the four lines still undisclosed.
Four lines remain undisclosed, and they are the deal.
The production runtime rate for agent traffic, the MCP tool call unit, the completion pass through for model tokens, and the rate lock term: across the order forms we reviewed, the production runtime rate was blank in most first drafts.
And the completion pass through was the least understood line at every table it appeared on.
The blank line is not neutral.
A placeholder rate fills in later at Pay as you go, on the meter's terms, after the build has committed you. Buyers who insisted on a named, capped rate before signing paid materially less per action.
And the locked rate cut effective cost 20 to 35 percent against Pay as you go across the reviewed deals.
The budget method is bottom up or fiction. Model credit burn on peak volumes, actions per workflow at the 20 credit anchor, Voice at its premium, the agent to agent growth the architecture implies, then add the supervising seats at $125.
First order forms priced the seats clearly and left the consumption meter vague, which is exactly backwards from where the money is.
The anchors versus the blanks
| Line | Status | The budgeting posture |
|---|---|---|
| Flex Credits at $500 per 100,000 | Published anchor | The conversion rate every action estimate runs through |
| 20 credits per action, 30 for Voice | Published anchor | The per action cost, near 10 and 15 cents, that workflow models multiply |
| $125 per user per month add on | Published anchor, $150 regulated | The supervising seat layer, scoped to the cohort that touches agent workflows |
| Production runtime rate | Undisclosed, blank in most first drafts | Named and capped before signature, or it fills in at Pay as you go |
| MCP tool call unit | Undisclosed | Priced before external agents consume the tool surface at volume |
| Completion pass through | Undisclosed, least understood | The model token line demanded in writing: who pays, at what rate, with what cap |
An undisclosed line is a negotiation that has not happened yet. Salesforce not publishing a rate does not mean the rate is undecided; it means it is decided per deal, and the first draft's blank is an invitation to sign before the number exists.
The buyers who filled the blanks on their own terms set the market the stragglers will eventually pay.
The burn model, budgeting from the workflow up
The model that held across our reviews starts at the workflow: actions per task per agent design, tasks per day at target adoption, the Voice share at its 30 credit premium, and the agent to agent multiplication the headless architecture exists to produce.
Converted at the credit anchor and stress tested at peak volumes rather than averages.
The Headless 360 pillar carries the full meter mechanics, and the MCP tool call analysis prices the tool surface the undisclosed unit will eventually bill.
The forecast history says to pad upward: account team credit estimates ran 40 to 70 percent below realized consumption once agents went live, anchored on chat volume while the non interactive traffic compounded. The peak volume model is not conservative; against that record, it is realistic.
The Agentforce Flex Credits playbook
The credit forecasting worksheet, the buying model comparison, the capped rate clause language, and the undisclosed line checklist for the order form.
Get the white paper →The capped rate, the single strongest move
The reviewed deals split on one clause: a named production runtime rate, capped, locked for the term, versus a placeholder resolving to Pay as you go.
The locked cohort paid 20 to 35 percent less per action, and the mechanism is ordinary: the rate negotiated before the build starts is negotiated against alternatives, and the rate accepted after the agents ship is negotiated against sunk cost.
The Headless 360 versus Agentforce boundary decides which meters your traffic actually draws, and the Agentforce pricing guide carries the buying model detail underneath.
- Percentile standing for your exact deal size and industry, from real closed transactions
- Scenario simulation before the call: test alternative terms and see the financial impact of each
- A negotiation playbook, talking points, and a two page executive brief on day one
What the order form reviews showed, 2025 to 2026
Across the Agentforce and Data Cloud order forms Fredrik Filipsson reviewed in 2025 and 2026, the pattern was structural rather than accidental:
First order forms priced seats clearly and left the production consumption meter vague or empty.
Effective per action cost against Pay as you go for buyers who named and capped the rate before signing.
The completion pass through finding deserves its own line: the model token cost flowing through agent responses was the least understood item on every order form it appeared on.
And the least negotiated, which for a line that scales with reasoning model adoption is exactly the wrong item to leave ambient.
Who pays for the tokens, at what rate, with what cap, in writing, is the whole ask.
Your first five moves
- Budget from the anchors: workflow level burn at 20 credits per action, Voice at 30, converted at $500 per 100,000, stress tested at peaks.
- Name and cap the production runtime rate before signature; the blank fills in at Pay as you go otherwise.
- Demand the completion pass through in writing: who pays for model tokens, at what rate, with what ceiling.
- Price the MCP tool call unit before external agents consume it at volume, and the rate lock term alongside it.
- Scope the $125 seats to the supervising cohort, never the license population. The Salesforce practice runs the order form with you, blank by blank.
Frequently asked questions
How much does Salesforce Headless 360 cost?
There is no standalone production price list: the cost flows through the meters underneath, anchored at $500 per 100,000 Flex Credits, 20 credits per Agentforce action, 30 for Voice, and the $125 per user per month add on.
The realized cost is your modeled traffic times the rates you negotiate into the order form.
What Headless 360 pricing is still undisclosed?
Four lines: the production runtime rate for agent traffic, blank in most first draft order forms, the MCP tool call unit, the completion pass through for model tokens, and the rate lock term.
Each is decided per deal rather than published, which makes each a negotiation the first draft hopes you skip.
How should we budget for Headless 360 before prices are published?
Bottom up from the workflow: actions per task per agent design, tasks per day at target adoption, Voice at its premium, and the agent to agent growth the architecture implies, converted at the credit anchor and stress tested at peak volumes, with the supervising seats added at $125.
Account team estimates ran 40 to 70 percent low.
What is the completion pass through?
The model token cost flowing through agent responses, passed to the buyer on terms most order forms leave ambient.
It was the least understood line in every review, and it scales with reasoning model adoption, which makes the written answer, who pays, at what rate, with what cap, a mandatory order form item.
What is the single best move on a Headless 360 deal?
A named, capped production runtime rate locked for the term, negotiated before the build starts. The locked cohort paid 20 to 35 percent less per action than Pay as you go across our reviews, because rates negotiated against alternatives beat rates accepted against sunk cost, every time.
Does Headless 360 need its own budget line?
It needs its own model, running through shared meters: the credits, the add on seats, and the undisclosed lines all bill through the Agentforce and platform constructs the rest of the estate uses.
The budget discipline is per workflow burn modeling with the consumption terms pinned, the same as every agent surface on the platform.
