# IrisAgent — AI support agents that resolve chat, email, voice and text inside your existing helpdesk.

> Source: The Agents Index — https://theagentsindex.com/irisagent (structured, researched, re-verified)
> Facts last verified: 2026-09-24

IrisAgent is an AI customer-support platform whose grounded agents resolve chat, email, voice and SMS conversations inside an existing helpdesk such as Zendesk, Salesforce, Intercom or Freshworks. It adds agent assist, automated tagging and routing, conversation QA and ticket analytics on top of that helpdesk.

| Fact | Value |
| --- | --- |
| Website | https://irisagent.com |
| Price | from $500/month |
| Free tier | Yes |

Standing $0/mo plan, no credit card required. Includes IrisGPT chatbot with GenAI answers, CRM connectors, automated tagging and email support only (4 of the 35 listed features); agent handoff, Slack/Teams bot, voice, triaging, analytics and SSO are excluded.

## What it does

**Done-for-you service** · Partly — Normally you buy software you run yourself, but the Managed Resolution line has IrisAgent staff operate the AI on your helpdesk and bill only for conversations they resolve.

> We run IrisAgent on your existing helpdesk and resolve your support conversations end to end. You pay per resolution. No seats, no platform fee, nothing for your team to install.
> — https://irisagent.com/managed-resolution/ · Hero section, 'Outcome-based pricing' introduction

## How much it does unattended

**Runs autonomously** · Yes — Triggers can send an AI answer to the customer with no agent involved, and a resolution is only counted when no human handoff was needed; low confidence hands back to a person.

> A support conversation (chat, email, or phone call) is resolved when our AI answers it and no agent handoff is needed. If a human has to step in, it does not count as a resolution
> — https://irisagent.com/managed-resolution/ · 'What counts as a resolution' section

**Multi-agent** · Partly — One deployment carries several specialised agents (support, IT help desk, knowledge, ticket automation, QA), but the vendor never documents several agents coordinating on a single conversation.

> one private deployment powers customer support, internal IT and employee help desk, knowledge, and operations agents.
> — https://irisagent.com/on-premise-ai/ · 'What are on-premise, self-hosted enterprise AI agents?' bullet 'Agents across the enterprise'

**Agent permissions** · Partly — Autonomy is configurable per trigger: the same AI answer can be sent straight to the customer or posted as an internal note an agent approves, and enabling deflection everywhere is restricted to admins.

> This is a good first step when rolling out AI answers, since the agent reviews the answer and sends it in one click rather than the AI replying to the customer directly.
> — https://docs.irisagent.com/deploying-irisagent/Tickets · 'Help your agents' → 'Write private note' action

## Which models it runs on

**Claude** · Yes — Anthropic models are scored inside the production IrisAgent stack on its own leaderboard, and the vendor says production routing follows that board.

> Every model runs inside the same production IrisAgent stack. Same retrieval, same tools, same prompts. Only the model changes.
> — https://irisagent.com/support-model-arena/ · 'Four axes that matter in support' intro, above the leaderboard listing Claude Opus 5 and Claude Fable 5

**GPT** · Yes — OpenAI models appear in the same in-stack leaderboard, and the security FAQ names OpenAI-based models as one of the two GenAI options offered to customers.

> We have two options for GenAI models - private, in-house models or OpenAI-based models.
> — https://irisagent.com/security/ · Security FAQ, 'What models do you use for Generative AI? Where is the data sent?'

**Open models** · Partly — Open models are offered for the private, self-hosted deployment where weights run inside the buyer's own environment; the shared SaaS names only in-house or OpenAI-based models.

> IrisAgent Private runs open and private models inside your environment so prompts and responses never touch a third-party model provider.
> — https://irisagent.com/on-premise-ai/ · FAQ, 'Which AI models does it use, and can we use our own?'

**Model choice** · Partly — The vendor calls itself model-agnostic but picks the defaults from its own leaderboard and routes on the buyer's behalf, so model selection is a vendor-configured choice rather than a self-serve setting.

> Yes. IrisAgent is model-agnostic. We pick defaults from this board and can route by customer, channel, and latency budget.
> — https://irisagent.com/support-model-arena/ · FAQ, 'Can I use a different model with IrisAgent?'

## Where you use it

**In a browser** · Yes — There is a hosted dashboard where triggers, automation and analytics are configured, plus embeddable web chat and search widgets for customers.

> Go to **Automate** → **Case Triggers** on the [IrisAgent dashboard](https://web.irisagent.com/triggers) and click **Create New Trigger**.
> — https://docs.irisagent.com/deploying-irisagent/Tickets · 'Create a trigger', step 1

## Whose machine it runs on

**Self-hosted** · Partly — Self-hosting is a separately purchased line, IrisAgent Private, installed in your data centre or your own AWS, Azure or GCP account; the standard Free and Standard plans are vendor-hosted SaaS.

> Run IrisAgent inside your own environment, on-premise or in your own cloud. Full data sovereignty, zero egress, and audit-ready SOC 2, HIPAA, and GDPR alignment.
> — https://irisagent.com/pricing/ · Add-ons section, 'Private, Self-Hosted AI' card

**Open source** · No — The vendor calls its grounding engine proprietary and publishes only docs and an OpenAPI spec publicly; no product source or self-build path is offered.

> Our proprietary engine grounds every response in your data. No made-up answers, ever.
> — https://irisagent.com/pricing/ · 'Why teams choose IrisAgent', Hallucination-Free AI card

## What it costs to run

**How it meters** · Yes — The published Standard plan is a monthly fee that scales with query volume, while the Managed Resolution line charges per resolved conversation with no seat or platform fee.

> $500/mo Starting at · scales with query volume
> — https://irisagent.com/pricing/ · Plan comparison table, Standard column header

**Free tier** · Yes — A standing $0 plan covers the IrisGPT chatbot, CRM connectors, automated tagging and email support, and sign-up needs no credit card or sales call.

> The Free plan includes IrisGPT Chatbot with GenAI answers, CRM connectors, automated tagging, and email support. It's perfect for startups looking to get started with AI-powered support.
> — https://irisagent.com/pricing/ · Pricing FAQ, "What's included in the Free plan?"

**API access** · Yes — A documented public REST API covers asking, summarising, tagging, sentiment, ingestion and search, with an OpenAPI spec; keys are issued by an IrisAgent contact rather than self-serve.

> The IrisAgent API lets you use IrisAgent as a service from your own application. You can power intelligent conversations and search, and index and connect your internal and external knowledge articles, tickets
> — https://docs.irisagent.com/api-reference/introduction · 'Welcome' section of the API Overview

**Bring your own key** · Partly — Bringing your own approved models is offered only with the private deployment, where the buyer hosts the weights alongside the platform; no BYO key option is published for the SaaS plans.

> You can run the models we recommend and tune for accuracy, or bring your own approved models and host them alongside the platform. Either way the model weights and the data stay in your control.
> — https://irisagent.com/on-premise-ai/ · FAQ, 'Which AI models does it use, and can we use our own?'

_Not established: MCP server_

## Buying it for a team

**A company can buy it** · Yes — An Enterprise tier is sold through sales with custom deployment, SLAs and an account manager, and the vendor accepts ACH and wire transfers for it.

> We accept all major credit cards, ACH bank transfers, and wire transfers for Enterprise plans. Annual billing is available with additional savings.
> — https://irisagent.com/pricing/ · Pricing FAQ, 'What payment methods do you accept?'

**Seat model** · No — The published plans meter query volume rather than agent seats, and the managed line states outright that there are no per-seat fees, so no seat floor or ceiling is published.

> ✓ Pay per resolved conversation ✓ No per-seat fees ✓ No platform fee ✓ No charge when we hand off to your team
> — https://irisagent.com/managed-resolution/ · 'Outcome-based pricing' bullet list

**Admin controls** · Partly — Admins control whether AI answers reach customers at all and scope them with trigger conditions, and role-based access is described for the private deployment; no general policy-engine documentation was found.

> Access to the IrisAgent dashboard. Enabling deflection across all tickets is admin-only.
> — https://docs.irisagent.com/deploying-irisagent/Tickets · Prerequisites, item 3

**Audit log** · Partly — A full audit trail on every interaction is promised for the private deployment; for the hosted plans only trigger activity counts over 1, 7 or 14 days are documented.

> Full audit trail on every interaction, with role-based access controls and guardrails that keep regulated decisions with your team.
> — https://irisagent.com/on-premise-ai/ · 'Audit-ready answers your examiners can trust' section, second bullet

**Single sign-on** · Partly — Sign-in is always federated through Google, Microsoft, Salesforce or Okta OIDC, but the plan table puts the SSO / SAML row on Enterprise only, so the two vendor pages disagree on tiering.

> We do not require or host any login credentials. We authenticate through Google, Microsoft, Salesforce, and Okta.
> — https://docs.irisagent.com/security-and-compliance/Single-Sign-On-Options · Introduction section

_Not established: Pooled budget_

## What happens to your code

**Opt out of training** · Yes — The vendor states customer data grounds answers only for that customer and never trains anyone else's models, and generative AI can be switched off or kept on private in-house models entirely.

> Your data is used to ground answers for your own customers and is never used to train models for anyone else.
> — https://irisagent.com/accuracy/ · 'Security and trust, certified' section

**Data residency** · Partly — Region and jurisdiction are a buyer choice only on the private deployment, which runs in your own VPC or data centre; for the shared SaaS the vendor names no hosting region.

> Runs in your VPC, your region
> — https://irisagent.com/on-premise-ai/ · 'Your own cloud (VPC)' deployment card, first bullet

**Getting out** · Partly — Cancellation is any-time and all data is deleted on account closure, but no customer-facing data export path is documented and the API only covers upload and deletion of content.

> You can cancel your account any time and we delete all your data after you close your account.
> — https://irisagent.com/security/ · Security FAQ, 'What happens to my data if I choose to terminate the agreement with IrisAgent?'

**Certifications** · Yes — The vendor claims SOC 2 Type II certification plus HIPAA, GDPR, CCPA and PCI DSS alignment, with an air-gapped option for regulated buyers; ISO 27001 is not mentioned.

> Every AI customer conversation IrisAgent powers (chat, voice, email, and agent copilot) is handled under enterprise security and compliance controls.
> — https://irisagent.com/security/ · 'Certifications and compliance' section, badges SOC 2 Type II / HIPAA / GDPR / CCPA / PCI DSS

## Provenance

Every fact above comes from the vendor. Nothing here is independently corroborated yet.

## Coverage elsewhere

None recorded. Every fact above comes from the vendor and is not independently corroborated yet.

## What changed

- 2026-09-24 — Capabilities · Commercial terms · Free tier
- 2026-09-03 — Capabilities · Commercial terms · Free tier · Price

## Correcting this record

Anything marked "Not established" is a question we have not answered — it is not a No. If you represent IrisAgent, or anything here is wrong or out of date, you can supply it with a source: the operations this site accepts from an agent are listed at https://theagentsindex.com/.well-known/agents.json — that catalogue needs no key to read, and each entry states its own auth.

Supplying evidence is free. Nothing that can be bought here changes what a field says, whether a gap is closed, or how this record is scored.