IrisAgent

What it does

Done-for-you service

PartlyBoth models exist: you can license the software yourself, or take Managed Resolution where IrisAgent's team runs it on your helpdesk and bills per resolved conversation.We run IrisAgent on your existing helpdesk and resolve your support conversations end to end. You pay per resolution.irisagent.com/managed-resolution ↗

How much it does unattended

Runs autonomously

YesThe vendor claims autonomous end-to-end resolution of the majority of volume, with escalation rather than a guess when confidence drops below its threshold.IrisAgent autonomously resolves 50%+ of tickets, and many teams see 60%+ of conversations resolved end to end as grounding and workflows mature.irisagent.com/faq ↗

Multi-agent

PartlyOne 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.irisagent.com/on-premise-ai ↗

Agent permissions

YesPermissions are set per intent: guardrails and escalation rules are written as plain-English procedures, and auto-send only happens above a confidence threshold you choose.encode guardrails and escalation rules as Smart Operating Procedures written in plain English, with no code requiredirisagent.com/faq ↗

Which models it runs on

Claude

NoThe security FAQ enumerates exactly two model options for the hosted product, in-house private models or OpenAI, and never names Anthropic as a model IrisAgent runs.We have two options for GenAI models - private, in-house models or OpenAI-based models.irisagent.com/security ↗

GPT

YesOpenAI is one of the two model options offered, it is opt-in per approved data, and it can be disabled entirely in favour of the private models.For the latter, only the data that you approve will be sent to OpenAI for processing.irisagent.com/security ↗

Open models

YesFine-tuned open-weight models handle the cheapest routing tier on IrisAgent's own infrastructure, and private deployments can run open models inside your boundary.These go to smaller, faster models: fine-tuned open-source models that we host on our own infrastructure.irisagent.com/blog/how-our-multi-llm-en… ↗

Model choice

PartlyOn the hosted product the vendor routes each query by intent and complexity and states the model choice is not the buyer's; only private deployments let you bring approved models.The model choice is hidden from the buyer; the outcome is the metricirisagent.com/llm-customer-support ↗

Where you use it

In a browser

YesA customisable embeddable widget runs on any website, alongside a hosted dashboard for configuration, conversations and analytics.IrisAgent's customizable chatbot can be deployed on any website to answer support questions.docs.irisagent.com/deploying-irisagent/… ↗

Whose machine it runs on

Self-hosted

YesIrisAgent Private installs the whole stack in your data centre or your own VPC, including an air-gapped variant with no outbound connectivity; it is software only, no appliance.Regulated teams can deploy the full IrisAgent stack for AI customer conversations inside their own data center or their own cloud (VPC on AWS, Azure,…irisagent.com/security ↗

Open source

NoThe 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.irisagent.com/pricing ↗

What it costs to run

Free tier

YesA permanent Free plan carries the chatbot, CRM connectors, tagging and email support with no card and no sales call required.The Free plan includes IrisGPT Chatbot with GenAI answers, CRM connectors, automated tagging, and email support.irisagent.com/pricing ↗

API access

YesA documented public REST API covers ask, summary, tag, sentiment and content ingest, though keys are issued by your IrisAgent contact rather than self-serve, and no plan tier is named.The IrisAgent API lets you use IrisAgent as a service from your own application.docs.irisagent.com/api-reference/introd… ↗

Bring your own key

PartlyBringing your own models or your own cloud only holds in the private deployment line; hosted customers choose between IrisAgent's private models and OpenAI, not their own keys.Run open or private models in your environment, or bring your own approved models. Weights and data stay in your control.irisagent.com/on-premise-ai ↗

Not established: How it meters, MCP server

Buying it for a team

A company can buy it

YesAn Enterprise tier is sold through procurement with ACH and wire payment, dedicated account management, training and SLA guarantees rather than an individual card checkout.We accept all major credit cards, ACH bank transfers, and wire transfers for Enterprise plans.irisagent.com/pricing ↗

Seat model

NoNothing is priced per seat: the entry paid tier is a flat monthly figure and the managed lane bills per resolution, and no seat floor or ceiling is published on any plan.There are no seat fees, no platform fee, and no charge for conversations we hand off to your team.irisagent.com/managed-resolution ↗

Admin controls

YesAdmins constrain the agent through guardrails on sensitive operations, per-intent escalation rules, confidence thresholds and role-based access scoped per brand or region.Built-in guardrails ensure critical operations like identity verification, high-value refunds, and compliance-sensitive requests are handled securely.irisagent.com/smart-operating-procedures ↗

Audit log

PartlyA per-answer trail showing the response served and its source is claimed for enterprise deployments, but logging of administrative configuration changes is never asserted for IrisAgent itself.Single sign-on through your identity provider, role-based access control, encryption in transit and at rest, defined data residency and retention,…irisagent.com/enterprise-ai-chatbot ↗

Single sign-on

PartlyEvery account signs in through Google, Microsoft, Salesforce or Okta OIDC with no local passwords, but the plan comparison lists SSO / SAML as an Enterprise-only row, so SAML federation reads as gated.We do not require or host any login credentials. We authenticate through Google, Microsoft, Salesforce, and Okta.docs.irisagent.com/security-and-complia… ↗

What happens to your code

Opt out of training

YesNo opt-out is needed because the default is retrieval-based grounding on your own content, and generative processing through OpenAI can be switched off entirely.Your data is not used to train shared or third-party foundation models.irisagent.com/faq ↗

Data residency

PartlyNo default jurisdiction is published; residency is negotiated during enterprise onboarding, or fixed by deploying IrisAgent Private into your own VPC and region.We scope residency requirements as part of enterprise onboarding rather than applying a one-size-fits-all default.irisagent.com/faq ↗

Getting out

PartlyCancellation any time with deletion of all your data is committed, but no self-serve export or data-portability path is published; nothing is prepaid as credits, so expiry does not arise.You can cancel your account any time and we delete all your data after you close your account.irisagent.com/security ↗

Certifications

YesSOC 2 Type II, GDPR, CCPA, PCI DSS alignment and HIPAA readiness are all claimed, with reports available on request; ISO 27001 is absent from that enumerated list.IrisAgent is SOC 2 Type II certified and GDPR compliant. Data is encrypted at rest and in transit, and we operate on an enterprise-grade security…irisagent.com/faq ↗