Yardstick Research tear-sheet / AI sales cohort

Methodology · how we score · rubric weights in plain sight · vendors received this sheet seven days before publication and could flag factual errors, never rankings

Lorikeet

Identity

• Founded: 2023, public launch October 2024 [THIRD-PARTY — https://www.linkedin.com/company/lorikeetcx and https://www.startupdaily.net/topic/funding/ai-customer-support-startup-lorikeet-has-now-raised-75-million-in-just-10-months/]
• HQ: New York, NY on LinkedIn; operations split between Sydney, Australia and the United States per vendor about page (split signal) [THIRD-PARTY — https://www.linkedin.com/company/lorikeetcx and https://www.lorikeetcx.ai/about]
• Legal entity: [UNKNOWN] Originally incorporated as Optech per startup-press coverage; current registered legal name not surfaced in this research session
• Funding: ~US$51.3M total raised across three rounds in roughly ten months — US$7.3M Seed (Oct 2024, Square Peg lead), A$14M (~US$9M) Seed extension (Feb 2025, Blackbird lead), US$35M Series A (Aug 6, 2025, QED Investors lead). Reported "$200M+" valuation at Series A close. [THIRD-PARTY — https://www.prnewswire.com/news-releases/lorikeet-raises-35-million-usd-series-a-to-deliver-ai-powered-universal-customer-concierges-able-to-solve-real-challenges-not-just-chatbots-302522821.html and https://www.startupdaily.net/topic/funding/ai-customer-support-startup-lorikeet-has-now-raised-75-million-in-just-10-months/]
• Headcount: ~90 employees per LinkedIn (May 2026) [THIRD-PARTY — https://www.linkedin.com/company/lorikeetcx]
• Founders: Steve Hind (CEO, ex-Stripe + Watershed product), Jamie Hall (ex-Google Brain, named author on Google's Meena and LaMDA papers) [VENDOR-CLAIMED — https://www.lorikeetcx.ai/about]
• Recent news (last 12 months):
• January 29, 2026 — Coach launched: AI co-worker that reviews 100% of conversations, surfaces root-cause shifts in support metrics, proposes and deploys fixes with approval. Reference customer: HotDoc (telehealth, 13M+ patients). [VENDOR-CLAIMED — https://www.prnewswire.com/news-releases/lorikeet-launches-coach-the-first-ai-that-tells-you-why-your-support-metrics-moveand-fixes-the-problem-302673653.html]
• December 2025 / May 2026 — Voice 2.0: multilingual (Spanish, Mandarin), identity verification, virtual-card creation, concurrent calls at scale. GiveCard reference: 200,000 SNAP recipients reached, deployed in <48h. [VENDOR-CLAIMED — https://www.lorikeetcx.ai/blog/lorikeet-voice-2-0]
• October 9, 2025 — "Team of Agents" launch with voice integration: multi-agent orchestration that can call logistics providers, text doctors, email hotels. Named customers: Airwallex, Flex, Linktree, Arbor, Eucalyptus. [VENDOR-CLAIMED — https://www.prnewswire.com/news-releases/lorikeet-launches-industry-first-team-of-agents-with-voice-integration-to-deliver-end-to-end-customer-problem-resolution-302579515.html]
• August 6, 2025 — Series A: US$35M, QED Investors lead; vendor cited "10x revenue growth since October public launch." [VENDOR-CLAIMED for revenue, THIRD-PARTY for round]
• Latent risk — No G2 / Capterra / Trustpilot footprint; SOC 2 + ISO 27001 evidence pack gated to Scale tier and above on trust center; underlying LLM provider not publicly disclosed. [THIRD-PARTY — https://www.eesel.ai/blog/lorikeet-ai-alternatives (competitor source — weight accordingly)]
• Archetype: Deterministic-workflow AI customer concierge for regulated, multi-step support — fintech, healthtech, insurance, regulated marketplace.

Total score: 68.75 / 100

• Stage fit:
• Foundation (<40 readiness): no — entry price ($18k/year) and sales-led onboarding (no free tier, no published simulation mode) rule out Foundation-stage buyers without a real ticket volume to justify the budget.
• Pilot (40–59): yes — Start tier at $1,500/month with transparent per-resolution credit pricing is exactly sized for a Series A/B fintech or healthtech running a single-workflow pilot on KYC, refunds, or claim status.
• Scale (60–79): yes — Scale tier at $4,000/month unlocks trust-center access (SOC 2 + ISO 27001 evidence) and lower per-resolution unit costs. Sized for Series B/C running 50k–500k tickets/year across multiple workflows.
• Optimization (80+): conditional — Enterprise tier delivers dedicated engineering + HIPAA BAA + custom security review, but the pricing is opaque and the customer mix tilts away from Fortune-500 IDN-tier procurement. A buyer at >5M tickets/year or in Fortune-100 healthcare procurement should expect a custom-scoped pilot rather than the published Scale envelope.
• One-line verdict: Best deterministic-workflow concierge in the CX cohort for regulated fintech and healthtech at Pilot and Scale stage; bounded weak spots are public evidence gaps (undisclosed LLM, gated trust center, no third-party review footprint) rather than architectural ones.

Headline numbers

Dimension scores

Penalty pre-score (from dossier §6.5): Integration penalty: none. Scale penalty: soft (vendor does not publish total annual conversation volume or named Fortune-500 / enterprise IDN-tier customers; customer mix tilts to growth-stage fintech / healthtech). Already reflected in the published headline score of 68.

Pricing detail

Source: https://www.lorikeetcx.ai/pricing (vendor's own pricing page; cross-checked against assembled.com teardown). All [VENDOR-CLAIMED]; annual-billing rate shown.

• Start — $1,500/month annual ($18,000/year): 18,000 credits/year. Chat / email / SMS at 0.95 credits per resolution, voice at 1.50 credits per resolution, ticket routing+tagging at 0.15 credits each, automated QA at 0.20 credits each. "We only charge for successfully resolved tickets."
• Scale — $4,000/month annual ($48,000/year): 48,000 credits/year. Chat / email / SMS at 0.80 credits, voice at 1.20 credits, routing+tagging at 0.10 credits, automated QA at 0.20 credits. Trust-center evidence pack (SOC 2 Type II + ISO 27001) opens at this tier.
• Enterprise — Custom: Customized credit allocation across all service types. Dedicated engineering team, "custom team across entire life of contract," custom security reviews. HIPAA BAA availability gated to this tier.

No free tier and no published self-serve trial. Entry path is a sales-led pilot.

Integrations

Source: https://www.lorikeetcx.ai/integrations (cross-checked against assembled.com 2026 teardown).

• Help-desk / CRM (native connectors): Intercom, Salesforce, Zendesk, Zendesk Messaging, Front, HubSpot, Help Scout — 7 platforms. Vendor describes the Salesforce connector as "execute support workflows inside Salesforce for consistent, auditable outcomes" and Zendesk as "turn Zendesk tickets into completed work — classification, action, and closure."
• Knowledge bases: Confluence, Guru, Google Docs, Notion.
• Action connectors: Stripe (payments + customer records in real-time), Shopify (order + customer info), SendGrid (transactional email + engagement tracking).
• Communication / messaging: Slack (support requests + internal workflows), Twilio (SMS + voice infrastructure).
• Telephony / CCaaS: Genesys Cloud, Amazon Connect, Five9, Dialpad, Vonage, AirCall — 6 named connectors. This breadth is unusual at Lorikeet's stage and is what makes the voice channel claim credible.
• Developer surface: Lorikeet API, Lorikeet SDK, MCP support ("extend agent capabilities with custom tools and data sources").
• Webhook support: [UNKNOWN] not explicitly listed on the public integrations page; likely present given the SDK but not confirmed in this research session.

Editorial assessment

Lorikeet is the cohort's clearest "deterministic workflow" alternative to free-form LLM concierges like Sierra, Decagon, and Maven AGI. The architectural bet is that for fintech, healthtech, insurance, and regulated marketplace workflows — refund processing, KYC follow-up, claim status, prescription follow-up, SNAP-emergency disbursement — the wrong-answer cost is large enough that every decision path the customer touches should be an editable, auditable workflow node rather than a generated reply. The named-customer evidence clusters tightly around that thesis: Airwallex, Flex, Linktree, Arbor, Eucalyptus, HotDoc, GiveCard, Taptap Send all fit the regulated, multi-step-workflow profile, and the head-to-head wins against Decagon at Flex and the GiveCard SNAP deployment (200,000 recipients reached in under 48 hours) are concrete signals the architecture holds under regulated-workload pressure.

The pricing is unusually honest for the cohort. Start ($1,500/month, $18k/year) and Scale ($4,000/month, $48k/year) publish per-resolution unit economics with credit rates a buyer can plug into a back-of-envelope cost-per-resolved-ticket model without filing an RFP. Voice resolutions price at roughly $1.00–$1.25 each, chat / email / SMS at roughly $0.67–$0.79 each. That arithmetic puts Lorikeet at chat parity with Intercom Fin's $0.99 flat resolution rate at the low end and slightly under at Scale, while the outcome-aligned credit model means buyers only burn credits on successfully resolved tickets. The pricing transparency drops at Enterprise tier — which is the tier most regulated-industry buyers will land on — and that opacity is the one consistent third-party critique.

The bounded weak spots are evidence gaps, not architectural gaps. The underlying LLM is not publicly disclosed in any source we surveyed. No accuracy benchmark is published. No contractual hallucination guarantee exists (Glia is the only cohort vendor with one). No public penetration-test report. No public sub-processor list. The SOC 2 Type II and ISO 27001 evidence pack is gated behind the Scale-tier paywall on the trust center. None of these are deal-breakers for a growth-stage fintech buyer running a pilot, but enterprise IDN-tier healthcare procurement teams asking for the full audit pack will need to extract it during the sales cycle rather than self-serve at trust.lorikeetcx.ai. The third-party review-site footprint is also thin: no G2, no Capterra, no Trustpilot — buyers anchor on vendor references and a sales-led pilot.

Best fit is Series A through Series C fintechs and healthtechs running 50k–500k support conversations per year, where the workflow-completion outcome (refund successfully processed, KYC successfully passed, prescription successfully followed up) is the SLA, not deflection rate. The cohort leaders (Glia, Maven AGI, Sierra) edge out Lorikeet on enterprise breadth and audit-evidence depth; Lorikeet edges them on per-resolution pricing transparency and on the explicit deterministic-workflow architecture for buyers who value that specifically. When to revisit: if Lorikeet opens the trust center evidence pack to all tiers, publishes a named LLM provider and version policy, publishes an accuracy benchmark, or lands a named Fortune-500 health-system reference, the score moves up meaningfully against the cohort.

Best for

• Stage: Pilot and Scale. Conditional fit for Optimization at Enterprise tier; not a Foundation-stage fit.
• Company profile: Series A through Series C fintech, healthtech, insurance, or regulated marketplace; 50k–500k support conversations per year; in-house engineering team capable of integrating against an API + SDK + MCP surface.
• Industry fit: Strong for fintech, healthtech, insurance, regulated marketplace. Moderate for B2B SaaS with complex workflows. Weak fit for low-volume Foundation-stage buyers (entry price), Fortune-100 enterprise procurement without a custom-scoped pilot (scale envelope unproven), or buyers who require named-LLM disclosure today for AI governance review (Lorikeet does not publish this).
• Annual tooling budget: $18k–$48k+ on Start / Scale; custom at Enterprise.
• Skip if: You need (a) a contractual hallucination guarantee on a regulated channel (Glia), (b) named Fortune-500 enterprise IDN-tier references (Maven AGI, Sierra, Glia), (c) named-LLM disclosure for compliance review today (Lorikeet has not published this), (d) a self-serve simulation or test-before-deploy mode, or (e) third-party review-aggregate validation as a procurement gate (no G2 / Capterra / Trustpilot footprint as of May 2026).

Right-of-reply gaps

The following [UNKNOWN] items from the dossier should be put to Lorikeet during the fact-check window. The vendor is invited to fill any of these in; gaps left open are reported as such in the published tear sheet.

• Registered legal entity name and state of incorporation. The "Optech" naming surfaced in startup-press coverage; Lorikeet has not surfaced a canonical current registered entity in any public source we found.
• Underlying LLM provider and version. No public statement. Buyers with AI-vendor governance requirements need this for compliance review.
• Published accuracy or hallucination-rate benchmark. Vendor positions deterministic-workflow architecture as the structural answer; a measured benchmark would strengthen the dimension-3 score above 3/4.
• SOC 2 Type II audit date and auditor name. Vendor confirms SOC 2 + ISO 27001 via Vanta but does not publish dates on the public surface.
• Public penetration-test summary (latest test date + finding count + remediation status). Not currently surfaced.
• Public sub-processor list. Required for many regulated-industry procurement reviews.
• Data residency options. US / EU / AU availability not stated on the public surface.
• HIPAA BAA scope. Confirmed as available at Enterprise tier; the BAA's published scope and any clinical-workflow caveats are not on the public surface.
• Total annual conversation volume processed across the customer base. Vendor publishes named-customer wins; the aggregate volume number would directly relax the scale penalty.
• Webhook support specification. API + SDK + MCP are listed; webhooks not explicitly called out on the integrations page.
• Median time-to-first-resolution-in-production across the customer base. GiveCard's "under 48 hours" is the fast end; a published median would directly inform the time-to-value dimension.
• Brand-voice fine-tuning + content-approval workflow + audit-trail export. Workflow architecture provides the substrate; a published spec on how buyers configure brand-voice control and approval workflows would directly inform the dimension-7 score.
• Confirmation or denial of the a16z "AI Spending Report" ranking the vendor cited; we did not verify against the underlying a16z report in this session.

Sources

Lorikeet first-party: - https://www.lorikeetcx.ai/ - https://www.lorikeetcx.ai/about - https://www.lorikeetcx.ai/pricing - https://www.lorikeetcx.ai/integrations - https://www.lorikeetcx.ai/articles/best-ai-customer-support-software - https://www.lorikeetcx.ai/blog/lorikeet-voice-2-0 - https://www.lorikeetcx.ai/blog/lorikeet-secures-5-million-in-funding-to-empower-cx-teams-with-first-ai-agent-that-offers-human-quality-support-at-scale - https://trust.lorikeetcx.ai/

Press releases / investor: - https://www.prnewswire.com/news-releases/lorikeet-secures-5-million-in-funding-to-empower-cx-teams-with-first-ai-agent-that-offers-human-quality-support-at-scale-302266748.html - https://www.prnewswire.com/news-releases/lorikeet-raises-35-million-usd-series-a-to-deliver-ai-powered-universal-customer-concierges-able-to-solve-real-challenges-not-just-chatbots-302522821.html - https://www.prnewswire.com/news-releases/lorikeet-launches-industry-first-team-of-agents-with-voice-integration-to-deliver-end-to-end-customer-problem-resolution-302579515.html - https://www.prnewswire.com/news-releases/lorikeet-launches-coach-the-first-ai-that-tells-you-why-your-support-metrics-moveand-fixes-the-problem-302673653.html - https://www.crunchbase.com/organization/lorikeet (anonymous fetch 403; cited via search snippet) - https://www.crunchbase.com/funding_round/lorikeet-seed--bad97b9a - https://www.startupdaily.net/topic/funding/ai-customer-support-startup-lorikeet-has-now-raised-75-million-in-just-10-months/ - https://www.linkedin.com/company/lorikeetcx

Third-party reviews / teardowns: - https://www.assembled.com/blog/ai-chat-agents-customer-support - https://topaiagent.ai/review/lorikeet-review/ - https://www.eesel.ai/blog/lorikeet-ai-alternatives (competitor source — weight accordingly) - https://aiagentsdirectory.com/agent/lorikeet - https://www.g2.com/products/lorikeet/competitors/alternatives - https://coldiq.com/tools/lorikeet - https://justcall.io/ai-agent-directory/lorikeet/ - https://tracxn.com/d/companies/lorikeet/