Calculator
What does per-resolution AI billing cost at your volume?
Plug in your monthly AI-handled conversations. See the annual bill on every major vendor next to the cost of running the same volume on your own LLM keys, with every rate editable and every default sourced.
Your volume
Everything updates live. Vendor rates are editable below if you have negotiated numbers.
Conversations the AI resolves end-to-end. Per-resolution vendors bill on roughly this number.
Default $0.02 is a generous all-in figure for Gemini 2.5 Flash: multi-turn, RAG context, embeddings, retries. Raw single-pass list price is nearer $0.001.
A flat per-site fee; LLM tokens are billed to you directly by your provider. Business tier (dedicated infrastructure, SLA) is custom-quoted and not modelled here.
Vendor rates (edit if you've negotiated)
Defaults are published July 2026 rates; sources are linked below the calculator. Freddy bills per session, and one conversation can consume more than one session, so its real total skews higher.
OpenAgent, all-in
$1,236/year
$103/month: $100 LLM API + $3 platform fee. Unlimited agent seats, no per-resolution meter.
Annual bill at 5,000 conversations/month
Vendor rows exclude per-seat helpdesk fees (Intercom from $74/agent/mo, Zendesk from $55/agent/mo), so their real totals are higher than shown. Not every conversation becomes a billable outcome on every vendor; see the pricing comparison article for the counting rules.
Keep the 30× for yourself
Paste your Gemini, OpenAI, or Anthropic key and the widget is live. The platform fee is flat; the token bill stays between you and your provider.
Where the defaults come from
Every default rate is a published or widely reported list price, checked July 15, 2026:
- Intercom Fin, $0.99: the per-outcome price on intercom.com/pricing, billed per resolution, procedure handoff, or disqualification, with a 50-outcome monthly minimum.
- Zendesk AI, $1.50: the committed-volume rate reported by independent pricing teardowns (pay-as-you-go runs about $2.00); overages auto-bill at the full rate since January 2026.
- Salesforce Agentforce, $2.00: the per-conversation price on salesforce.com/agentforce/pricing. Flex Credits ($0.10 per action) can price a multi-step answer higher.
- Freshdesk Freddy AI, $0.49: the per-session rate; sessions are not conversations, and a single conversation can consume several sessions.
- LLM cost, $0.02 per conversation: a generous all-in estimate for Gemini 2.5 Flash ($0.30/M input, $2.50/M output tokens on the Gemini API price list) covering multi-turn exchanges, RAG context, embeddings, and retries. Single-pass list-price math is nearer $0.001.
What the calculator deliberately leaves out: per-seat helpdesk fees on the vendor side (they only make the gap wider), the custom-quoted Business tier (dedicated infrastructure and SLA), and negotiated enterprise discounts (paste yours into the rate fields to model them). Ada and Decagon are excluded because neither publishes a rate; see the pricing comparison for their reported contract sizes.