The first thing to understand about Amazon Connect pricing is that it breaks the model most contact centre buyers are used to. There are no per-seat licences, no minimum monthly fees, no long-term commitments, and no charge for peak capacity — you pay for what you use. Content was rephrased for compliance with licensing restrictions. [Source: AWS Amazon Connect Pricing]
That's liberating if your volume is variable, and a little disorienting if you're used to a predictable per-agent invoice. Let's break down exactly what you pay for.
Important: Exact prices vary by AWS region and change over time. The figures below are illustrative to explain the structure of the pricing. Always check the official AWS pricing page for current rates in your region before budgeting.
The Core Pricing Components
An Amazon Connect bill is made up of several separate usage-based charges. The main ones:
| Component | How it's charged | What it covers |
|---|---|---|
| Voice usage (service) | Per minute | The Amazon Connect service handling the call |
| Telephony (carrier) | Per minute, separately | The actual inbound/outbound carrier cost — this is on top of the service minute |
| Phone numbers | Per number, per day | Each DID or toll-free number you claim |
| Chat | Per message or per interaction | Web/app chat sessions |
| Tasks | Per task | Work items tracked in Connect |
| AI features | Included via Unlimited AI (tied to channel usage) | Contact Lens, Amazon Q, self-service, agent assist & supervisor AI — see below |
| Storage | Standard AWS rates | Call recordings and logs stored in S3 |
The key insight: the headline "per minute" figure you see quoted is usually just the service minute. The telephony minute (the carrier charge) is billed separately, and phone numbers carry a daily charge. To understand your real cost, you add these together.
Rule of thumb: Real voice cost per minute ≈ Connect service minute + telephony minute. Then add daily number charges and storage. Budget on the total, not the headline rate.
The Unlimited AI Model — A Big Deal
This is the part that genuinely sets recent Amazon Connect pricing apart, and it's easy to miss. Rather than metering every AI feature separately, Amazon Connect now offers unlimited use of its AI capabilities, with the cost tied only to your underlying channel usage — not to how much AI you consume. AWS describes it as providing unlimited use of the AI capabilities that power self-service, agent assistance, and supervisor experiences, so you can optimise every step of the journey without cost-driven compromises. Content was rephrased for compliance with licensing restrictions. [Source: AWS Amazon Connect Pricing]
In AWS's own framing, this is "all-you-can-eat" AI pricing: first-party AI across all channels, priced against channel usage rather than AI consumption. Content was rephrased for compliance with licensing restrictions. [Source: AWS Contact Center Blog]
What's covered under Unlimited AI includes capabilities such as:
- Contact Lens — conversational analytics, sentiment, and summarisation
- Agent assist and Amazon Q — real-time guidance and generative answers for agents
- Self-service AI — the AI powering automated customer interactions
- Supervisor experiences — AI-driven oversight and insights
Why it matters: under a metered model, teams often ration AI — turning off Contact Lens on some queues, limiting summarisation — to control cost. Unlimited AI removes that trade-off: you can apply AI everywhere without watching a per-use meter, and your AI cost moves only with your channel (voice/chat) volume. For a business planning to lean heavily on AI, this can materially change the economics versus rivals that charge per AI feature or per conversation.
Two honest caveats: First, "unlimited AI" doesn't mean free — the cost is bundled into channel usage, so higher voice/chat volume still drives the bill. Second, exact inclusions, availability, and how this is packaged can vary by region and evolve over time. Confirm what's covered for your account and region on the official AWS pricing page before you budget.
A Worked Example
Let's make it concrete with an illustrative small operation — a 5-person team taking around 50 inbound calls a day, averaging 3 minutes each, on one local number. Using representative (not official) rates purely to show the method:
| Item | Rough calculation | Indicative monthly |
|---|---|---|
| Inbound voice (service + telephony) | 50 calls × 3 min × ~22 working days × combined per-min rate | ~£50-90 |
| Phone number | 1 number × daily charge × 30 days | ~£2 |
| Recordings storage | S3 storage of call recordings | A few £ |
| AI features | Included via Unlimited AI (no separate per-use meter) | Included |
| Indicative total | ~£60-100/month |
Compare that to traditional CCaaS at roughly £15-30 per seat per month plus line rental, hardware, and feature add-ons, and for this profile Connect usually comes out lower — the more so if volume is seasonal, because you're not paying for idle seats in the quiet months. We go deeper on this scenario in Is Amazon Connect Suitable for Micro Businesses?
Pay-As-You-Go vs Per-Seat: Which Wins?
Neither model is universally cheaper — it depends on your shape of demand:
- Pay-as-you-go (Connect) wins when your volume is variable or seasonal, you're growing, you want no upfront commitment, or you have quiet periods where per-seat licences would sit idle.
- Per-seat can be competitive when your volume is steady and predictable, agents are on the phone near-constantly, and you value a fixed, forecastable invoice over usage flexibility.
The strategic advantage of usage-based pricing is that cost scales with activity, not headcount. When AI handles more of your volume, you don't keep paying for seats that are no longer needed — a point that matters more every year as AI absorbs routine call volume.
Hidden Costs to Watch
The "gotchas" that catch teams out when the first bill arrives:
- Telephony is separate. The biggest surprise. The carrier per-minute charge is on top of the Connect service minute — always add both.
- Phone numbers add up. A daily charge per number is trivial for one number, but scales if you claim many (e.g. local numbers per region).
- Storage grows forever. Call recordings and logs accumulate in S3. Set retention policies so you're not paying to store years of audio you'll never use.
- AI is bundled, not free. Under Unlimited AI, AI features aren't a separate per-use meter — but the cost is bundled into channel usage, so heavy voice/chat volume still drives the bill. Don't assume "unlimited" means "no cost".
- Outbound and international. Outbound calling and international destinations carry different (often higher) telephony rates.
The advertised-rate trap: A quoted service rate is not your all-in cost. Once telephony, numbers, and storage are included, your true cost per interaction is higher. (AI is bundled via Unlimited AI rather than added per-use, but channel volume still drives the total.) Always budget on modelled total cost per interaction.
How to Estimate Your Own Cost
- Get your call data. Daily call volume, average handle time, inbound vs outbound split.
- Look up current regional rates on the official AWS pricing page for both the service minute and telephony minute.
- Add the components: (calls × minutes × [service + telephony rate]) + (numbers × daily charge × 30) + storage. AI is bundled via Unlimited AI, so you don't add it per-feature — but remember it's reflected in the channel rate.
- Use the AWS Pricing Calculator to model it, and add a buffer for growth.
- Compare like-for-like against per-seat quotes — include their hidden costs (line rental, add-ons) too.
Frequently Asked Questions
How does Amazon Connect pricing work?
Amazon Connect uses a pay-as-you-go model with no per-seat licences, minimums, or long-term commitments. You pay for usage: per-minute voice service charges plus separate per-minute telephony, a daily charge per phone number, and separate usage charges for chat, tasks, and AI features.
Is Amazon Connect cheaper than per-seat contact centre software?
It can be, especially for variable or seasonal volumes, because you pay for usage rather than fixed seats — and because AI is bundled under the Unlimited AI model rather than charged per feature, which can beat rivals that meter every AI capability. For steady high-volume operations the comparison is closer — model your real usage including telephony, numbers, and storage rather than relying on the headline per-minute rate.
What are the hidden costs of Amazon Connect?
The main overlooked costs are telephony/carrier charges (separate from the Connect service minute), phone number daily charges, S3 storage for recordings and logs, and add-on AI features. Model total cost per interaction, not just the base per-minute rate.
Conclusion
Amazon Connect pricing rewards you for paying only for what you use — a genuine advantage for variable, seasonal, or growing operations. The catch is that the bill is assembled from several separate usage components, so the headline per-minute rate understates your real cost. Add the telephony minute, the number charges, and storage, and model it against your actual call data. The standout, though, is the Unlimited AI model: AI capabilities are bundled with channel usage rather than metered per feature, which removes the "should we ration AI to control cost?" question that dogs metered rivals. Just remember unlimited doesn't mean free — channel volume still drives the total.
Deciding whether Connect fits your business? Read Is Amazon Connect Suitable for Micro Businesses? and the Complete Amazon Connect Feature Guide.