Lorikeet

The integrations page documents named connectors across ticketing and messaging (Zendesk, Zendesk Messaging, Front, Help Scout, Slack), CRM (Salesforce, including running workflows inside it) and revenue tooling, with tools set up as APIs or webhooks so the agent reads data and takes actions in arbitrary internal systems, scoped least-privilege with short-lived single-operation tokens and HMAC-signed inbound webhooks.

Sourcelorikeetcx.ai/integrationsread 2026-09-05

Lorikeet executes multi-step work against documented procedure: it publishes that whether it is a three-step form or a forty-step refund flow, the agent executes with precision, following the customer's own standard operating procedures rather than improvising, and the product surface names the workflow primitives it runs on, including workflow matching, ticket triage, gather, get status, choice and route based on problem, with tool calls such as retrieving recent transactions shown inline. Lorikeet frames this as specialized agents collaborating in real time across channels to deliver end-to-end resolutions, and the login surface is a workflows console.

Sourcelorikeetcx.airead 2026-09-05

Lorikeet grounds its agent on the customer's own documented procedures rather than on retrieval alone. It states plainly that it follows your SOPs, taps into your systems and handles your highest-stakes workflows, and has said that rather than relying solely on retrieval-augmented generation its architecture goes deeper by following the procedures themselves, so the agent neither merely retrieves nor invents process. Help centers, reference materials and internal tools supply the surrounding corpus, so the grounding is a maintained structure the customer authors and owns.

Sourcelorikeetcx.airead 2026-09-05

Runtime guardrails detect sensitive topics and escalate to the customer's team before the agent proceeds, outbound checks block the agent from disclosing withheld fields or acting outside policy independently of the model, tools are scoped least-privilege with per-user elevation required for dangerous actions, and the integration is red-teamed against prompt injection and scope escalation before go-live. Humans author the procedures the agent follows and handoffs arrive with the investigation complete.

Sourcelorikeetcx.ai/product/trustread 2026-09-05

SOC 2 Type II, ISO 27001:2022, HIPAA and GDPR, all independently verified and published on a public Vanta trust center at trust.lorikeetcx.ai with reports downloadable under NDA and refreshed annually, plus signed BAAs for healthcare customers. Named controls: automatic PII redaction, minimum-necessary PHI handling, TLS 1.3 in transit and AES-256 at rest, no public internet access to production, row-level-security tenant isolation, SSO with hardware-key MFA, and third-party penetration testing.

Sourcelorikeetcx.ai/product/trustread 2026-09-05

Every customer interaction, model choice and action is tracked and stated to be fully explainable rather than a black box, with Lorikeet's integration guidance describing every tool call, elevation request and reasoning step logged with timestamps and replayable for a regulator examination, and Coach quality-scoring every conversation, human or AI. That reconstructs the decision path rather than only logging the conversation.

Sourcelorikeetcx.ai/product/trustread 2026-09-05

Lorikeet names a customer memory but does not describe its shape. It states that customer memory maintains context across every touchpoint, and that a customer who starts in chat does not repeat themselves on a call because the channels run on one engine, so persistence spans channels rather than sitting inside a single session. No scope, lifetime, retention window or deletion control for that memory is documented.

Sourcelorikeetcx.airead 2026-09-05

Production runs on Google Cloud Platform in a private VPC with no public internet access and row-level-security tenant isolation, US primary hosting, and data storage residency the customer can take in Australia or the EU. Lorikeet's content-marketing articles name the US, Australia and the UK, but the product trust page states Australia and the EU.

Sourcelorikeetcx.ai/product/trustread 2026-09-05

Configuration is authored per customer from their own standard operating procedures, which is the deliberate product position rather than a gap: Lorikeet's stated test is that if you can explain to a human agent how to resolve a ticket, the agent can handle it, so time to value depends on the customer's procedure documentation.

No template library, prebuilt agent catalog or marketplace is documented, and the specialized agents in the Team of Agents architecture are internal collaborators rather than assets a buyer browses and adopts. Lorikeet also runs a separate property at lorikeet.tools, branded Toolshed, and what it holds is not described.

Sourcelorikeetcx.airead 2026-09-05

Five channels run on one engine plus a programmatic surface: the homepage names phone, SMS, chat, email and WhatsApp, with an SDK listed alongside them as a delivery surface, and Lorikeet describes specialized agents collaborating in real time across chat, email, voice and SMS to deliver end-to-end resolutions. Voice runs at sub-second latency on the same engine, so a customer who starts in chat does not repeat themselves on a call, and every channel inherits the same guardrails and logging.

Sourcelorikeetcx.airead 2026-09-05

The trust page names OpenAI, Anthropic and Baseten among core subprocessors, so several model providers are in play, held to zero-data-retention agreements with no fine-tuning on customer data, and Lorikeet describes routing dynamically between providers by task. The routing is Lorikeet's own and the model choice is tracked rather than chosen; no customer-facing selection or bring-your-own-model control is documented.

Sourcelorikeetcx.ai/product/trustread 2026-09-05

A developer documentation site exists at docs.lorikeetcx.ai covering a quickstart and API reference, and Lorikeet's integration guidance describes scoped Bearer-credential access and HMAC-signed inbound webhooks the customer's systems call. Lorikeet also publishes public HTTP endpoints that external AI assistants can call to query its support agent and book demos. The documentation site sits behind a customer access code, so the platform API and its scope are open only to customers, and no SDK is advertised.

Sourcedocs.lorikeetcx.ai/guides/quickstartread 2026-09-05

Lorikeet tests on both sides of go-live, with a readable result on each. Before launch, a pre-go-live simulation suite lets a security or compliance team validate agent behavior before unsupervised resolution begins.

After launch, Coach ships as a named product that reviews every ticket rather than a sample, whether handled by humans, AI or both, surfaces why performance is trending up or down, proposes fixes and, with approval, implements them, with automated quality assurance and thematic analysis and segmentation and export. Coach is sold to Lorikeet customers and standalone, is metered per ticket, and reached general availability in January 2026. That is a gate on the release path and a scored optimization loop after it, with the customer's own agent as what is tested.

Sourcelorikeetcx.airead 2026-09-05

Work happens programmatically inside the customer's systems through controlled integrations, tapping into Zendesk, Stripe and internal APIs, with tool calls such as retrieving recent transactions shown as named workflow steps. No browser control, hosted session or computer use is documented. Its public HTTP endpoints for external AI assistants are Lorikeet exposing itself to be called rather than operating an interface.

Sourcelorikeetcx.airead 2026-09-05