Agentforce
Salesforce · Platform
Salesforce’s platform to build and run AI agents inside Salesforce, priced by consumption, by conversation or per user.
- Published price
- How it is billed
- Does it train AI on your data?
- Certificates and audit reports
ECONOMICS LENS
A per-seat plan, a minute of voice and a resolved conversation describe different units. Inspect the billing scope before comparing prices.
Separate potential capacity from converted financial value. Compare full operating costs only after recording the rates, review burden and quality assumptions.
Worked example: every number is an assumption, not a customer result or current provider quote. Edits stay in this browser session. Assumption links and prepared enquiry drafts can contain numeric inputs; source notes are excluded from those links. Shared or older input links are sent to hosting when opened.
Negative values are retained. Time released is not cash saved; realization requires measured outcomes, a conversion plan and attributable cost changes.
Baseline = annual workload × avoidable cost per unit. Gross headroom = baseline × improvement. Converted value = headroom × conversion. Recurring net = converted value − incremental annual cost. Year-one net also subtracts setup. Match the workload, period and accepted quality; do not count the same saving twice.
Source model: supplied Expanded Research & Economics workbook, 27 September 2026, Global Cases and Scenario Inputs. The public-processing example retains its negative low case. Your assumptions define this range; measured customer savings require a baseline and observed results.
Use a dated quote or provider rate card for model costs and your own workload, review and cost observations. Record product, plan, region, date and assumptions before using the result for a decision. No provider rate is fetched automatically.
Editable hypothetical example · not provider pricing or measured performance. Use this monthly worksheet to test your own quote and quality assumptions.
/ accepted outcome
accepted outcomes / month
5,000 tasks × 80% = 4,000 accepted outcomes per month. Higher entered costs raise unit cost; more accepted outcomes spread the same cost over more results.
Formula: entered monthly costs ÷ (monthly tasks × acceptance % ÷ 100). The numerator includes only the costs you enter. Add applicable licence, usage, integration, human review and rework costs yourself; this is not automatically a fully loaded estimate or a vendor billing model. Capacity released is not realized financial value.
Inputs stay in this browser tab’s session, including reload and language changes. They are not added to the page URL. Continuing prepares an unsent enquiry draft.
84 objects with a recorded billing basis. Missing prices remain open research points.
Salesforce · Platform
Salesforce’s platform to build and run AI agents inside Salesforce, priced by consumption, by conversation or per user.
Zendesk · Product
AI agents inside Zendesk that answer customer requests; Zendesk bills them per “automated resolution”.
Intercom · Product
An AI agent that answers customer requests and is billed per outcome; Intercom states Fin works with “no seats required”.
Salesforce · Product
Salesforce’s AI for customer service. The service AI page leads with: “AI in the flow of work and your unmetered employee agent, all in one.”
NiCE (Cognigy GmbH) · Product
A platform to build and run AI agents for contact centres by phone, chat and messaging, plus a copilot for service staff.
Freshworks · Product
AI in Freshworks’ help desk: a self-service agent for customers and a copilot that helps service staff answer faster.
PolyAI Ltd · Product
AI agents that answer customer phone calls and can continue the same conversation in chat, SMS and messaging.
Genesys · Product
A contact-centre platform with AI agents (“Agentic Virtual Agent”), copilots for agents and supervisors, and AI Studio.
Augury · Product
Sensors and AI that monitor machine condition and predict failures, plus AI for process optimisation.