Decagon
An enterprise AI agent for customer service across chat, voice and email, configured with natural-language procedures. By Decagon.
100% of reviewers who mention it are positive on “resolves customer questions on its own” (10 of 10; 95% interval 72%–100%). 5 reviews report wrong or invented answers.
What reviewers say
10 public reviews about the AI, from G2 (via AWS Marketplace). Each share counts the reviews that mention the point.
Read with care. All of these reviews were posted within six weeks in 2026, with similar wording and mostly top ratings. That pattern often comes from a review-collection campaign, though none is labelled as incentivized.
Does the job
100%72%–100%10 of 10
Accuracy
0 of 5too few to rateHandoff to a person
1 of 3too few to rateCost
–too few to rateSetup and upkeep
1 of 5too few to ratePraise
“Sometimes the AI needs additional context or a human agent to step in, so better handling of edge cases and smoother handoffs would make the support workflow more efficient.”
“One thing I think Decagon could improve is the accuracy and consistency of responses in some situations.”
“The initial setup was fairly straightforward, and the platform was intuitive enough to get started without a lot of friction.”
“If I notice any problems or mistakes in company policies, I find that the resolutions provided by Decagon might not be proper and correct.”
“It may make mistakes or misunderstand unusual customer problems , so a human support agents may be still be needed”
“It handles a large volume of routine requests quickly while giving our team the flexibility to step in when human attention is needed.”
Complaints
“Sometimes the AI needs additional context or a human agent to step in, so better handling of edge cases and smoother handoffs would make the support workflow more efficient.”
“One thing I think Decagon could improve is the accuracy and consistency of responses in some situations.”
“The initial setup was fairly straightforward, and the platform was intuitive enough to get started without a lot of friction.”
“If I notice any problems or mistakes in company policies, I find that the resolutions provided by Decagon might not be proper and correct.”
“It may make mistakes or misunderstand unusual customer problems , so a human support agents may be still be needed”
“Sometimes the responses need to be reviewed or corrected by a human, especially when the conversation requires deeper context.”
What it does
Decagon sells an AI agent for customer service over chat, voice and email1. Teams define the agent's workflows in natural language as Agent Operating Procedures (AOPs), and Duet can draft AOPs from uploaded conversation transcripts1,2. Engineering teams can version workflows with Git-based tracking2.
- Channels
- Chat, Voice, Email
- Works with
- Decagon names Salesforce, Zendesk and Intercom for tickets and routing, Kustomer for knowledge sync, and Amazon Connect and RingCentral for voice3.
Handing a conversation to a person
As Decagon documents it. Reviewers' experience of handoffs is in the shares above.
Controls
Safety and review features, as Decagon documents them.
- Guardrails for brand voice, escalations and hallucinations across chat, voice and email2
- Simulations test agent behavior before production and with each update; Duet can generate tests, and checkpoints confirm the agent used the right actions, data and tools4
- Users can see why a test passed or failed, schedule recurring runs, and send alerts to tools such as PagerDuty4
- CI/CD integration to test every agent version, and Git-based version tracking for workflows4,2
- Watchtower for always-on QA and Experiments for live A/B tests1
- Tracing of the agent's reasoning at any point in a conversation2
Setup
Decagon says most enterprises go live "in weeks rather than months" and that a forward-deployed team is included with every engagement5.
What Decagon claims
The vendor's own figures. Shown as claims; never used in a result.
Not found in public sources
We looked for these and could not find them.
- Public prices
- What a human agent receives at handoff
- An approval workflow for changes to AOPs
- A formal definition of a billed resolution or conversation
Billing
By contractNot published; sold by contract. Decagon offers per-conversation pricing (a fixed rate per incoming conversation) or per-resolution pricing (a higher fixed rate per fully resolved conversation, no charge for escalations), and says most customers choose per-conversation6,5.
What counts: Decagon frames resolution as a conversation resolved by the AI agent without escalating to a human; it publishes no formal billing definition6.
Ask what counts as a billable outcome, and whether there is a minimum commitment.
Reviews found
- About the AI
- 10
- Sources
- 1
- Collected
- Oct 8, 2026
Where we looked
- Trustpilot decagon.ai: 0 reviews
- AWS Marketplace reviews page (prodview-pnbnafhjefhhy): 29 external G2 reviews, 10 shown (page parameter ignored); all 10 used
- G2: 403
- Capterra: 403
- noizz.io: 429, not read
- Reddit (2 searches): no threads found
Also in AI agents
Markdown copy: decagon.md. Something wrong? See corrections. Get Support AI is not affiliated with Decagon.