At What Actions Per Conversation Does Flex Credits Beat the $2 Conversation SKU?

Run the list-price arithmetic before the AE call. Salesforce publishes both meters on the same pricing page: Flex Credits at $500 per 100,000 credits, and Conversations at $2 per conversation. On the same page, Salesforce's own worked examples price a one-action use case at 20 Flex Credits / $0.10 and a voice call example at 120 Flex Credits / $0.15. Source: salesforce.com/agentforce/pricing.

That means a standard action at list is $0.10. Divide $2 by $0.10 and the break-even is 20 standard actions per conversation. Below 20, Flex Credits win at list. Above 20, the flat Conversation SKU is cheaper at list. Salesforce does not publish that crossover as a board slide. Competitors already do. This page is the Director+ worksheet, not a vendor calculator.

This is not the A1E/E1E prompt-meter question. It is not the Flex Credit overage forecast. It is not Digital Wallet's missing kill switch. It is not BU chargeback. It is the buying-model fork Salesforce itself puts on the pricing FAQ: Flex Credits versus Conversations, and you cannot run both in one org.

The two list meters, in buyer language

Meter A, Flex Credits. Pay per action as you scale. Public list: $500 per 100k credits. Salesforce says Flex Credits are fungible across Actions, Prompts, Translations, and Voice Actions. Digital Wallet is the near-real-time consumption view. Buying-model options include Pre-Purchase, Pre-Commit, and PayGo. Source: salesforce.com/agentforce/pricing.

Meter B, Conversations. Flat $2 per conversation at USD list, regardless of how many actions sit inside the conversation. Optimized in Salesforce's own FAQ language for external-facing customer agents. Conversation consumption is viewable in Digital Wallet, but Flex Credits get the more granular usage data. Same pricing URL.

Org rule that changes the board slide. Salesforce's FAQ: Flex Credits and Conversations will not be supported in the same org. You pick a model. You do not blend by agent. Same URL.

The break-even worksheet RevOps should run

Use list prices only as the planning floor. Contracted rates beat list. Do not hard-code a forever price into the forecast. Re-check the rate card and the order form every quarter.

1. Sample 30 to 50 real conversations (or sandbox replays) for each agent type you plan to put in production.
2. Count discrete actions per conversation the way Digital Wallet will bill them: lookups, updates, tool calls, retrievals, voice turns. Do not count messages if Salesforce bills actions.
3. Compute mean and p90 actions per conversation by agent type.
4. Apply the list break-even: if mean actions stay under 20 standard actions, Flex Credits are cheaper at list. If mean or p90 sits above 20, Conversations can win on unit cost and caps the long messy case.
5. Re-run the same math with voice multipliers if voice is in scope. Salesforce's pricing FAQ says Agentforce Voice usage types and multipliers differ; see the Agentforce Rate Card. Source: salesforce.com/agentforce/pricing.
6. Weight by monthly volume. A FAQ agent at 1 to 3 actions can make Flex Credits look brilliant while an SDR sequence at 30-plus actions flips the org onto Conversations if that volume dominates.
7. Add the commitment trap. Unused Flex Credits do not roll over into subsequent subscription terms. Salesforce says so on the same FAQ. Oversizing Pre-Purchase to feel safe is how finance funds waste.
8. Name the owner. RevOps owns the action-count sample and the break-even table. Procurement owns the SKU swap language if you are leaving Conversations. Finance owns the month-end match to Digital Wallet.

Worked list-price examples (illustrative, not a quote)

These use Salesforce's published list arithmetic only. They are not your contract.

• FAQ / knowledge lookup: 1 to 3 actions / conversation → Flex Credits $0.10 to $0.30 vs Conversations $2.00 → Flex Credits cheaper at list.
• Case triage / routing: 5 to 9 actions → $0.50 to $0.90 vs $2.00 → Flex Credits.
• Field scheduling: ~6 actions → ~$0.60 vs $2.00 → Flex Credits.
• Break-even: 20 actions → $2.00 vs $2.00 → tie at list.
• Heavy SDR / multi-tool sequence: 30-plus actions → $3.00-plus vs $2.00 → Conversations cheaper at list.

Salesforce's own pricing-page examples already sit on the cheap side of the line (1, 2, 3, 5 action shapes). The SERP gap is the crossover itself: consultants and calculator sites are ranking on when $2 beats Flex Credits, and our live decoder set answers adjacent meters without publishing the 20-action worksheet.

What the sibling decoders already cover (do not duplicate)

• Which SKU path burns Flex Credits only vs Einstein Requests plus Flex Credits: SKU decoder
• How to forecast Flex Credits when overages bill in arrears: Flex Credits forecast
• Digital Wallet alerts without a hard stop: Digital Wallet
• Who audits the meter and what audit rights to write in: Who audits the meter
• How to chargeback Flex Credits when Digital Wallet is org-level: BU chargeback

This page is the buying-model fork before those pages matter.

What to put in writing before the next AE call

1. Observed mean and p90 actions per conversation by agent type (sample method and date).
2. Whether voice is in scope, and which rate-card multiplier the AE confirms in writing.
3. Which buying model you are asking for (Flex Credits Pre-Purchase / Pre-Commit / PayGo, or Conversations), knowing the org cannot mix.
4. Whether you are swapping existing Conversation SKUs to Flex Credits (Salesforce FAQ says a full SKU swap is required).
5. Rollover rule restated: unused Flex Credits do not carry. Size the commitment to observed usage, not hope.
6. Named Digital Wallet alert thresholds and a kill rule that is not wait for the invoice (sibling decoder).

Related: SKU path · Flex Credits forecast · Digital Wallet · BU chargeback · Core, Advanced, and Max editions