Shopping Agents That Take Payments: Sierra vs Fin vs Agentforce vs Decagon

Sierra announced a Stripe partnership on October 20 that lets a customer talk to an AI agent, pay, and receive a receipt on the same call. The first public case is Cedar, a healthcare billing company whose agent is called Kora. Sierra reports that 71.3 percent of calls that reach a payment start end with a successful payment.

That moves support agents from answering questions to closing transactions. It also makes platform selection harder, because the four vendors below differ more in billing structure than in feature lists, and feature grids hide that difference.

Four billing models, four company sizes

Start with how the money works, because it is the hardest thing to change later.

PlatformBilling modelPublic pricingPayment inside the conversation
SierraPer resolutionNo public rate card. Third-party estimate: platform floor around $150K per year, implementation $50K to $200K, $1 to $2.50 per resolutionYes, Stripe integration announced October 20, 2026
Fin (formerly Intercom)Per resolution$0.99 per resolution; plans from $49 per month including 50 resolutionsNot stated in published material
AgentforcePer conversation or Flex CreditsCommonly reported at $2 per conversation, or $500 per 100,000 creditsOrders through Commerce, not payment inside the chat
DecagonEnterprise contractNo public rate card. Third-party estimate: median around $432,750 per yearNot stated in published material

The gap between Sierra and Fin is the point. One is a six-figure annual contract; the other is 99 cents per resolution. They do not serve the same buyer, and a feature comparison across both produces a misleading answer.

Sierra was reported at a $15.8 billion valuation in May 2026 and its customer base skews large enterprise. Fin’s parent Intercom was acquired by Salesforce, a deal completed September 10, which puts Fin and Agentforce inside the same organization.

What payment inside a call unlocks

The Stripe integration deserves separate treatment because it is the only capability here with a dated public launch.

The flow: the agent detects payment intent and switches into a secure mode. The customer enters card details on the phone keypad, the digits route directly to Stripe, and the agent returns to confirm the transaction and handle whatever comes next. Card numbers never pass through the conversational model.

For cross-border sellers, the relevant channel is the phone. Returns, price adjustments, subscription renewals, and B2B invoice settlement still land on voice lines in the US and Europe. Previously an agent could only transfer the caller to a human to collect. Now it can collect.

Published cases concentrate on medical billing. PCI scope, chargeback handling, and refund authority in a retail context are not detailed in the public material. If your category has a high return rate, ask specifically.

Channels and commerce connectors

Sierra and Decagon emphasize phone plus website in public material. Fin is strongest in email, chat, and in-app messaging. Agentforce depends on Service Cloud and Commerce Cloud.

On commerce connectors: if your store runs on Shopify, Fin and Sierra both have existing integrations, while Agentforce and Decagon lean on Salesforce Commerce or custom work. If you already run Commerce Cloud, Agentforce has the lowest integration cost because it shares the data model.

There is no universal winner here. A company already committed to one platform pays far more to switch platforms than to switch agents.

Three questions before you buy

What are your support volume and order value? A store handling a few thousand sessions a month at tens of dollars per order should use outcome pricing, which is Fin’s model. Outcome pricing only pays off when each resolution carries real value, such as recovering a price difference or renewing a subscription.

Does the agent need to complete the transaction? If yes, Sierra is currently the only vendor with a first-hand public case. If no, skip that budget line entirely.

What is your existing platform? Shopify shops should look at Fin and Sierra first. Salesforce shops should look at Agentforce. Decagon-style enterprise contracts make sense mainly for custom stacks.

Two cautions

First, these prices reflect public information as of October 2026. Sierra and Decagon publish no rate card, so the figures above come from third-party estimates and quoted contracts will differ. In negotiation, pin down what counts as a resolution before arguing about the unit price.

Second, launching an agent is the start of the work. The holiday forecast cited earlier found retailers with brand agents growing holiday sales 6.2 percent against 3.9 percent without. That gap comes from whether the agent can read accurate product and order data, not from which vendor you pick. Data integration after selection takes longer than selection itself.

Related Articles

AI Traffic Control Compared 2026: How Cloudflare, Akamai and Vercel Classify Bots Differently

Cloudflare split AI traffic into Search, Agent and Training in July 2026. Akamai sorts the same traffic into training crawlers, search crawlers and fetchers. Vercel's AI bots managed ruleset covers training, search and user-generated fetches. Three vendors, nearly identical splits, different names and different defaults. A selection guide with a comparison table, an honest read on robots.txt, and a recommended setup for three kinds of store.