Salesforce’s new AI contact center platform is one of the most complete offerings the industry has seen. Here’s how to size it accurately so you can build a business case that actually holds up.
The customers who will get the most out of it are the ones who go in with an accurate cost model. The pricing structure has several layers that aren’t immediately obvious, and underestimating them early is the fastest way to kill an otherwise solid business case. Here’s how to build the number correctly.
Start with your seat license cost
Agentforce Contact Center sits on top of Salesforce Service Cloud, so your cost model always starts with the base license. The tiers stack like this:
| License layer | Per agent / month |
| Service Cloud Enterprise | $175 |
| Contact Center Add-on | $125 |
| Entry-level total | $300 |
| Full AI stack (top tier) | $625 |
Multiply by your headcount — that’s your baseline annual commitment before any AI activity runs. For a 200-seat center at mid-tier, you’re looking at roughly $1 million per year in seat licenses. That’s the floor your ROI case needs to clear.
Model your AI consumption separately
Salesforce prices AI activity through Flex Credits, where each action an AI agent takes costs approximately $0.10 (20 Flex Credits). Actions include retrieving a customer record, updating a CRM field, triggering a workflow, or sending a follow-up message.
A typical interaction generates 3 to 6 actions, meaning each AI-handled conversation runs $0.30 to $0.60 in consumption, which is separate from seat licenses.
Estimate your monthly AI spend:
Monthly interactions × containment rate × avg. actions × $0.10
Example: 100,000 interactions · 40% containment · 4 actions = $16,000/month in Flex Credits
Model this conservatively. Containment rates take time to optimize, and you’ll likely start lower than your target.
Understand what’s actually included
The top-tier package includes 1 million Flex Credits per year across your entire org — roughly 50,000 AI actions. At 4 actions per interaction, that covers about 12,500 fully contained conversations. A strong start for a pilot, but a fraction of monthly volume for an active center.
Once included credits are consumed, additional packs run $500 per 100,000 credits (5,000 actions). Build that cadence into your monthly run rate from day one.
Factor in voice and telephony
| Voice model | Cost | Works best when |
| Per-action | $0.15 / operation | High-volume, shorter calls |
| Per-minute | $0.30 / minute | Longer, more complex interactions |
Telephony platform costs are separate: $75–$200/user/mo with Amazon Connect, or ~$50/user/mo with bring-your-own telephony. Run both voice models against your actual call mix before committing.
Build in negotiation room (thoughtfully)
Multi-year agreements typically yield 15–30% off list. That’s a real lever, but use it to strengthen an already solid case rather than rescue an underbuilt one. When structuring your negotiation, focus on Flex Credit volume commitments, bundled telephony pricing, and ramp clauses if you’re phasing deployment over 6–12 months.
Validate the business case with three questions
- What is my cost per interaction today vs. with Agentforce?Divide total annual cost by expected interaction volume. AI containment should drive this number down — if it doesn’t, the case needs more work before you present it.
- What containment rate do I need to break even?Back-calculate the AI containment required to justify the incremental spend. If breakeven is 55% and your realistic starting point is 30%, build a clear ramp plan — not just an optimistic assumption.
- Have I included implementation costs?Deploying Agentforce well requires Salesforce expertise. Budget $150K–$500K for implementation and configuration — this line is consistently absent from first-draft models and consistently present in final invoices.
One more way to strengthen the economics
A significant portion of Flex Credit consumption in any deployment goes toward resolving data gaps — the AI agent making additional lookups, pulling from multiple systems, or triggering redundant workflows because customer context isn’t cleanly surfaced.
Platforms like DaVinci integrate directly with Agentforce to consolidate customer data before the AI agent touches it — reducing actions per interaction and improving containment rates at the same time. It’s the kind of complementary layer that helps an enterprise platform perform closer to its ceiling. If you’re building out your cost model, it’s worth including in the conversation.
Agentforce Contact Center represents a meaningful step forward for the industry. The organizations that will benefit most are the ones that go in clear-eyed about the full cost structure, set realistic milestones, and build the supporting infrastructure to get the most out of the platform. The technology can absolutely deliver — the key is giving it the right conditions to do so.
The author has 30 years of contact center experience and is currently building AI-powered contact center solutions.
