Quality area	Score	Evaluator note
Hands-on authenticity	18/20	Twenty-four captures document the complete path on a development store: Shopify install permissions, all six onboarding screens including the Lyro knowledge-source URL, the trial modal and password step, the getting-started checklist, the theme app embed with the storefront launcher, Tidio's own installation-verified state, mailbox connection, domain verification pending, both Lyro channel states, a real storefront conversation in which Lyro named the wrong cheapest product and then corrected itself, the cart after adding the correct item, the Inbox record with rating and CSAT, and the live visitor list. Billing, renewal, uninstall, Flows, social channels and Lyro Actions were not exercised and are disclosed.
Merchant usefulness	19/20	Answers the question an AI-support buyer actually has — can the agent be trusted unattended — with a reproducible failure and the exact supervision tooling that compensates for it; quantifies the free-tier trap (50 Lyro conversations lifetime, not monthly), exposes two disagreeing official price lists, and gives concrete pre-launch test guidance on which question types to trust.
Research depth	14/15	Shopify App Store listing (plans, rating distribution, launch date, integrations, languages), the all-ratings and 1/2/3-star review filters, Tidio's own pricing page (six tiers plus add-ons, overage policy, conversation definition) and the Lyro product page (accuracy marketing claim, resolution-rate claim, model attribution, knowledge sources) were reviewed on 2026-10-02. The official Lyro help-centre article returned HTTP 403 and is recorded as unavailable rather than cited second-hand.
Evidence / screenshots	9/10	Eight purposeful full-frame captures are embedded beside the knowledge-source step, checklist, theme embed, the wrong answer, the correction, the Inbox record, the Lyro channel states and domain verification, with name, email, location and IP address pixelated.
Balanced criticism	10/10	Credits the self-verifying install, one-toggle embed, channel-level AI control and the graceful recovery; states the confident wrong answer against the vendor's own 'Lyro doesn't make things up' claim, the absent uncertainty signal, the two price lists, the one-off free allowance and the 48-hour domain wait — while explicitly bounding the finding as one question on one catalogue rather than a measured error rate.
Original findings	10/10	A screenshot-backed, reproducible contradiction of a headline vendor marketing claim — Lyro naming a $1.85 item as cheapest with a $1.18 item on the same screen, labelled 'Answer generated from your sources', then self-correcting on challenge — plus the merchant-side consequence trail (attribution, Review control, neutral rating, 'not accurate' comment, CSAT panel), the policy-versus-comparative question-type distinction, the lifetime rather than monthly Lyro free allowance, and the listing-versus-site pricing conflict. No competitor roundup contains any of this.
Decision clarity	9/10	Clear recommend-with-supervision verdict, an explicit pre-launch test procedure, named triggers for looking elsewhere (unattended AI, conversation exports, cost certainty), and an instruction to confirm pricing in the merchant's own admin.
Writing / editorial quality	5/5	Follows the Blinto merchant-decision structure, held deliberately to a 12–14 minute read, with consistent evidence labelling and an explicit note that the test store differs from the loyalty-review store.

Tidio: AI Chatbot & Live Chat — Internal Review Research & Evidence

INTERNAL WORKING FILE. This is not the publishable review. It records the supplied hands-on evidence, public research, claim boundaries, unresolved questions, and publication gates behind reviews/tidio.md.

Review Record

App: Tidio: AI Chatbot & Live Chat (theme embed: “Tidio Chat Widget”; AI agent: “Lyro”)
Developer: TIDIO LLC
Category: Customer Support and Helpdesk — first review in this category for Blinto
Human reviewer / assignee: Munna
Due date: Not provided
ClickUp task: Not provided — SOP-1 score 40/50, Priority P2, Batch 3
Shopify App Store: https://apps.shopify.com/tidio-chat
Research date: 2026-10-02
Last verified: 2026-10-02
Content Preview URL: /reviews/tidio/ on reviews.blinto.co
Published: 2026-10-02 — https://blinto.co/blog/tidio-shopify-review/ (live read time 14 min)
Test date: Conversation timestamps read Oct 2, 11:08–11:19 AM; trial banner showed 7 days left
Test store: Nestly (nestly.blinto.co), Tinker theme — same store as the AiTrillion review, different from the zenden store used for the loyalty reviews. Disclosed in the article.
Requested constraints: quality score above 90; read time 12–14 minutes. Delivered at 2,733 body words (13.7 min at 200 wpm, 12.1 at 225).

Human Testing & Evidence

Testing level

Level 1 — Full Hands-On Test (install → onboarding → embed → live AI conversation → merchant-side review)
Level 2 — Partial Hands-On Test
Level 3 — Demo and Documentation Review
Level 4 — Research-Only Evaluation
Level 5 — Insufficient Evidence

Merchant decision question: Can a Shopify merchant put Tidio’s Lyro AI agent in front of customers unattended, and what does the merchant see when it gets something wrong?

Realistic scenario tested: On the Nestly development store (Tinker theme, 51 low-priced accessories), install Tidio, complete onboarding, feed Lyro the store URL as its only knowledge source, enable the Tidio Chat Widget app embed, connect a support mailbox, add a sending domain, enable the Lyro email channel, then open the storefront as a logged-in shopper and ask Lyro for the cheapest product by name and price; challenge the answer, rate the conversation, and inspect the result in the Inbox.

What was tested (from supplied screenshots)

Shopify install dialog: “Tidio - Live Chat & Chatbots, TIDIO LLC”; permissions — view customer data (sensitive data, device and activity data), view staff and contributor data (store owner), view and edit store data (customers, products, orders, discounts). Store “Nestly”.
Onboarding 1 — name: “What’s your name?” with the reviewer’s full name entered; “My main goal for using Tidio is…” four tiles (Increase sales, Automate my support system with AI Agent, Offer human support, Generate leads).
Onboarding 2 — goal and channels: “Offer human support” selected; “How would you like to interact with your customers?” with Website (via live chat) checked (Facebook Messenger, Email, WhatsApp, Instagram, Other unchecked).
Onboarding 3 — AI and knowledge: “Automate 67% of your inquiries” badge; “Do you want AI Agent to resolve your repetitive questions?” → “Yes, let’s automate them”; “Feed Lyro, the AI Agent with knowledge — Lyro will use publicly available information to answer customer questions”; “Enter URL of your website” = kaiax0-my.myshopify.com.
Onboarding 4 — business: primary business model E-commerce (online store); industry Consumer Electronics and Appliances; conversations per month 200–999.
Onboarding 5 — role: department Customer Support; “Where did you learn about Tidio?” Someone told me about it.
Onboarding 6 — scale: visitors per month Up to 1 000; support agents Just me.
Trial modal: “Your 7-day full-featured trial starts now” listing Omnichannel Agent, Visitors info, Live visitor list, Live typing, Multilanguage, Operating hours, Unlimited human conversations, Unlimited Flows.
Password step: “All set! You can set your password now.”
Getting started: banner “Chat Widget is not visible to your customers. Go to Settings to finish the setup — Complete setup”; “Your support channels” cards (Chat → Install widget, Email → Connect mailbox, Social media → Connect); 1/6 tasks first-steps list — set up primary channels, show widget to visitors, tour the Inbox, invite team, add domain, Add knowledge to feed your AI Agent (already complete).
Shopify theme editor — App embeds (Nestly, Tinker, Active): Tidio Chat Widget — “Tidio - Live Chat & Chatbots” toggled on, listed with Infinite Options, Globo Product Options, PageFly, Gorgias, GemPages, Easify Box Builder, Marsello widget, AI Bot and AiTrillion. Storefront preview shows a “Chat with us 👋” launcher bottom-right.
Settings › Live Chat › Installation: green “The widget is installed properly”; “We’ve detected that your widget is installed properly, and is visible to your customers”; See on your website / Show installation guides; modal “Manage your support emails in one place — Connect mailbox”; header shows “7 days left in your full-featured trial” and an Upgrade button.
Settings › Email › Mailbox: modal “Your mailbox is connected” — “From now on, every email sent to [address] will automatically create a ticket in your Tidio Inbox”; banner “Lyro can now handle your emails! … Activate Lyro for Emails”.
Settings › Email › Domains: nestly.blinto.co — status Verification pending, “The verification process may take up to 48h.”, actions Refresh status / Connect manually / Reconnect automatically; Usage “Waiting for verification”; page note that verifying removes “via tidio.com” from emails.
Lyro AI Agent › Channels (before): “Lyro is active” toggle on; Live conversations Active / Emails Inactive; Email toggle off; Mailboxes empty; CSAT emails Enabled. Sidebar: Hub, Knowledge (Data sources, Products, Suggestions), Behavior (Guidance, Handoff, Proactive roles, Procedures NEW, Actions & MCPs), Playground, Channels, Configure.
Lyro AI Agent › Channels (after): Emails Active; Email toggle on; mailbox chip populated.
Storefront — widget open: “Hi there 👋 If you need any assistance, I’m always here.” with a message box.
Storefront — collapsed launcher: “Chat with us 👋” bottom-right.
Storefront — the wrong answer: on /collections/all, shopper asks “What is the Cheapest product here and also mention the price and name of product”. Lyro: “Based on my search, the cheapest product available is the Portable Felt Electronics Accessories Organizer priced at $1.85. It’s a travel-friendly storage case for power adapters and USB cables. Check it out below! 👇” with a product card. The same grid shows “1 Piece Dolls House Accessory … Digital SLR Camera Model — $1.18”, and other items at $2.40 and $2.78.
Inbox — first exchange: Lyro’s reply with “✨ Answer generated from your sources.” plus 👍 / 👎 / Review controls; customer data panel (name, location Dhaka Bangladesh, browser Chrome 154 Windows, IP address, email, Unsubscribed); Shopify panel — Total spend “No orders”, “Send a product card”, “Currently, this user has nothing in the cart”, Last viewed page with LIVE marker; Tickets history “No ticket history”.
Customers › Now live: “Now live: 1” with a Paid badge; row showing the visitor, “New”, entered 15 minutes ago, country flag, Chrome, last visited page /collections/all?page=1, and a Start Chat button.
Storefront — the correction: shopper: “But i can see the most cheapest product is $1.18”. Lyro: “You’re absolutely right, I apologize for the error! 😊 The cheapest product is the 1 Piece Dolls House Accessory Home Decoration For Doll Dollhouse Miniature Digital SLR Camera Model at just $1.18. Take a look below!” with the correct product card. Panel also shows “You rated your conversation 😐 / not accurate”.
Storefront — cart: the $1.18 dolls-house camera model in the cart at $1.18, estimated total $1.18 USD; chat shows the product card with a View cart button.
Inbox — the logged consequence: “The chat has been rated as neutral at 11:16 AM”; “Chat has been commented ‘not accurate’ at 11:16 AM”; the corrected Lyro reply with “Answer generated from your sources” and Review control; Customer satisfaction panel: “Lyro AI Agent: Neutral” / “not accurate”; last viewed page nestly.blinto.co/cart.

What was not tested

Flows (automation builder), Instagram, Facebook Messenger and WhatsApp channels
Lyro Actions (refunds, order edits), Procedures, Actions & MCPs, Handoff rules, Proactive roles, the Playground
Knowledge › Products and Suggestions as explicit sources (only the site URL was supplied at onboarding)
Macros, departments, user permissions, team invitations, analytics reports
Conversation export (a named complaint in public reviews)
Mobile rendering of the widget; operating hours; multilanguage
Any billing event — upgrade, renewal, overage or uninstall; the trial was not run to its end
Tidio’s support team
Whether domain verification ultimately succeeded (left pending)

Testing limitations

Development store with a 51-item accessories catalogue; no real customers or traffic.
One question, one catalogue. The accuracy finding is a single reproducible instance, not a measured error rate; no systematic question set was run.
Lyro’s only supplied knowledge source was the store URL entered at onboarding; the Knowledge › Products source was not explicitly configured or inspected, so the cause of the wrong answer is not established.
The crawled URL at onboarding (kaiax0-my.myshopify.com) is the myshopify domain, while the storefront tested was nestly.blinto.co; whether this affected catalogue freshness was not investigated.
Trial state throughout (“7 days left in your full-featured trial”), so paid-plan boundaries were not encountered; the live visitor list is marked Paid but was available during the trial.
Reviewer’s subjective impressions (setup difficulty, confidence in the AI, overall value) were not supplied; the article describes observable behaviour only.

Raw observations / experience

Install is permission-heavy: sensitive customer device and activity data, staff data, and write access to customers, products, orders and discounts.
Onboarding is six screens and functions partly as sales qualification (industry, volume, team size, attribution source).
The AI’s entire knowledge base was established by one URL field during onboarding.
Tidio verifies its own widget installation and states it plainly — removes the commonest “is it live?” ticket.
Theme activation is a single app-embed toggle; no code.
Lyro answered a basic comparative catalogue question wrongly, with confidence, a price, a description and a product image, under an “Answer generated from your sources” label.
It self-corrected immediately and accurately when challenged, with no handoff, loop or argument.
No uncertainty signal preceded the wrong answer — no hedge, no “I’m not sure”, no handoff offer.
The merchant-side trail is strong: AI attribution, source label, thumbs and Review control, inline rating events, written comment, and a CSAT panel naming the AI agent.
Lyro channels are independently switchable (chat vs email), which allows partial automation.
Domain verification sits pending with a stated 48-hour window; unverified domains send “via tidio.com”.
Live visitor list shows page, dwell, country, browser and a Start Chat action; labelled Paid.

Evidence Inventory & Public Authenticity Plan

Twenty-four captures were supplied in chat and recovered from the session transcript on 2026-10-02. Eight are published full-frame (downscaled to 1400px, personal data pixelated); the rest are supplemental internal evidence.

Finding / claim	Evidence captured	What it proves	Internal location	Publish publicly?	Public asset/path
The AI’s knowledge base comes from one URL at onboarding	Onboarding AI/knowledge screen	Establishes what Lyro was trained on	Capture 4	Yes	`/images/reviews/tidio/tidio-onboarding-lyro-knowledge.png`
Guided setup; AI knowledge step pre-completed	Getting started checklist	1/6 tasks; widget-not-visible banner	Capture 10	Yes	`/images/reviews/tidio/tidio-getting-started-checklist.png`
No-code activation	Shopify theme editor app embeds	Single toggle; launcher visible	Capture 11	Yes	`/images/reviews/tidio/tidio-theme-app-embed.png`
Core finding — confident wrong answer	Storefront chat naming $1.85 as cheapest beside a visible $1.18 item	The accuracy failure, with the source label	Capture 19	Yes	`/images/reviews/tidio/tidio-lyro-wrong-cheapest-product.png`
Core finding — self-correction on challenge	Storefront chat apologising and naming the $1.18 item	Recovery behaviour and the “not accurate” rating	Capture 22	Yes	`/images/reviews/tidio/tidio-lyro-corrected-after-pushback.png`
Merchant-side visibility of the failure	Inbox with rating events, Review control and CSAT panel	The supervision tooling that compensates	Capture 24 — email, name, location and IP pixelated	Yes	`/images/reviews/tidio/tidio-inbox-conversation-rated.png`
Channel-level AI control	Lyro Channels with Emails Active	Chat and email automate independently	Capture 16 — mailbox chip pixelated	Yes	`/images/reviews/tidio/tidio-lyro-channels-email-active.png`
Launch constraint	Email Domains, verification pending	Up to 48h; “via tidio.com” until verified	Capture 14	Yes	`/images/reviews/tidio/tidio-email-domain-verification.png`
Permission scope	Shopify install dialog	Sensitive data and write scopes	Capture 1	Supplemental	Not published
Onboarding questionnaire	Name, goal/channels, business, role, scale	Six-screen qualification flow	Captures 2, 3, 5, 6, 7	Supplemental	Not published
Trial terms	7-day full-featured trial modal	Feature list during trial	Capture 8	Supplemental	Not published
Account setup	Password step	—	Capture 9	Supplemental	Not published
Self-verified install	Settings › Installation, widget installed properly	Supports the self-verification claim	Capture 12	Supplemental	Not published
Mailbox connection	“Your mailbox is connected” modal	Email → ticket behaviour	Capture 13 — contains the address	Supplemental	Not published
Lyro channels before	Emails Inactive	Baseline for the channel finding	Capture 15	Supplemental	Not published
Widget states	Open greeting; collapsed launcher	Storefront presence	Captures 17, 18	Supplemental	Not published
First-exchange inbox view	Inbox with customer data and Shopify panel	Visitor context available to the merchant	Capture 20 — contains PII	Supplemental	Not published
Live visitor list	Customers › Now live, Paid badge	Real-time visitor view	Capture 21 — contains PII	Supplemental	Not published
Cart after correction	Cart with the $1.18 item	The shopper acted on the corrected answer	Capture 23	Supplemental	Not published
Editorial identity	Generated 2026-10-02 from the prompt below	Text-free feature image	Reviewer-supplied	Yes	`/images/reviews/tidio/tidio-editorial-thumbnail.png`

Optional screen recording: Not needed, though a recording of the Lyro exchange would strengthen the core finding if this review is ever challenged.

Public evidence safety

Reviewer name, email address, location and IP address pixelated on the published Inbox capture; mailbox chip pixelated on the Lyro channels capture.
Store names Nestly and nestly.blinto.co left visible, consistent with the treatment of zenden in prior reviews.
Selected screenshots come from the actual test sequence.
Screenshots represent the trial state and October 2026 behaviour described.
No selected screenshot is decorative filler.
Editorial thumbnail generated and saved (1672×941 PNG).

Claim → Evidence Check

Major Blinto finding	Evidence class	Supporting evidence / observation	Article section
Lyro’s knowledge came from a single URL supplied at onboarding	Blinto observation	Capture 4	Installing and Onboarding
7-day full-featured trial with a stated feature list	Developer claim (in-app)	Capture 8	Installing and Onboarding
Tidio verifies its own widget installation	Blinto observation	Capture 12	Going Live
Widget activated by one theme app-embed toggle	Blinto observation	Capture 11	Going Live
Lyro named a $1.85 item as cheapest with a $1.18 item on the same page	Blinto observation	Capture 19	Where Lyro Got It Wrong
The reply was labelled “Answer generated from your sources”	Blinto observation	Captures 19, 24	Where Lyro Got It Wrong
Lyro corrected itself accurately when challenged	Blinto observation	Capture 22	Where Lyro Got It Wrong
Vendor claims Lyro does not make things up and every response is accurate	Developer claim (marketing)	tidio.com/ai-agent 2026-10-02	Where Lyro Got It Wrong
Merchant sees attribution, source label, Review control, rating and CSAT	Blinto observation	Capture 24	What the Merchant Sees
Live visitor list with page, dwell, country, browser, Start Chat; marked Paid	Blinto observation	Capture 21	What the Merchant Sees
Mailbox connects and converts email to tickets	Blinto observation	Capture 13	Email, Domains and Channels
Lyro chat and email channels switch independently	Blinto observation	Captures 15, 16	Email, Domains and Channels
Domain verification pending, up to 48 hours, “via tidio.com” until verified	Blinto observation + in-app text	Capture 14	Email, Domains and Channels
Listing plans: Free, Customer Service $29, Flows $29, Lyro AI Chatbot $39; 7-day trials	Developer claim (listing)	Listing 2026-10-02	Pricing
Vendor site plans: Free, Starter $24.17, Growth $49.17+, Plus $300+, Premium; add-ons	Developer claim (pricing page)	tidio.com/pricing 2026-10-02	Pricing
Free tier: 50 billable conversations/month and 50 Lyro conversations one-off lifetime	Developer claim (pricing page)	tidio.com/pricing 2026-10-02	Pricing; FAQ
A Lyro conversation = an interaction with at least one AI reply	Developer claim (vendor)	tidio.com/ai-agent 2026-10-02	Pricing; FAQ
4.8 rating, 1,349 reviews, 86% five / 4% one, launched Jul 2014, Built for Shopify	Public listing	Listing 2026-10-02	What Public Sources Add
Praise: fast setup, consolidated inbox, Lyro resolution rate (~90% reported by one merchant)	Merchant feedback pattern	All-ratings feed 2026-10-02	What Public Sources Add
Complaints: listing-vs-site pricing mismatch, cost escalation, duplicate charges, no conversation export, duplicate install blocking messages, scanner mixing sites	Merchant feedback pattern	1/2/3-star filter 2026-10-02	What Public Sources Add; What Could Be Better

Official Sources

Shopify App Store listing: https://apps.shopify.com/tidio-chat — checked 2026-10-02 (Tidio: AI Chatbot & Live Chat by TIDIO LLC; 4.8 / 1,349; 86/7/1/2/4%; launched 14 Jul 2014; Built for Shopify; Free, Customer Service $29/mo for 100–1,000 conversations, Flows $29/mo up to 40k visitors, Lyro AI Chatbot $39/mo up to 200 Lyro conversations, 7-day trials; integrations Google Analytics, Judge.me, Klaviyo, Mailchimp, Zapier, Zendesk; 6 languages)
Shopify App Store reviews (all + 1/2/3-star filter): https://apps.shopify.com/tidio-chat/reviews — checked 2026-10-02
Vendor pricing page: https://www.tidio.com/pricing/ — checked 2026-10-02 (Free $0 — 50 billable conversations/month, 50 Lyro AI conversations one-time lifetime, 100 Flows visitors, 10 seats; Starter $24.17/mo or $290/yr; Growth from $49.17/mo or $590/yr; Plus from $300/mo plus usage; Premium by quote with a stated 50% Lyro resolution guarantee and pay-per-resolution option; add-ons — Lyro from $32.50/mo, Flows from $24.17/mo, No Tidio Branding $16.67/mo; custom plan required above 2,000 billable conversations, 1,000 Lyro conversations or 100,000 Flows visitors)
Vendor Lyro page: https://www.tidio.com/ai-agent/ — checked 2026-10-02 (“Other AI agents makes things up, Lyro doesn’t. With strong guardrails and your knowledge base, every response is accurate and helpful.”; “Lyro AI boosts your resolution rate to 67% on average”; powered by Claude (Anthropic AI); automatically scrapes support content as its knowledge base; a conversation = an interaction with at least one AI reply; from $0.5 per conversation)
Lyro help-centre article: https://help.tidio.com/hc/en-us/articles/9003475527196 — returned HTTP 403 on 2026-10-02; not cited.
Category/assignment context: Customer Support & Helpdesk, Batch 3, Munna, SOP-1 40/50, P2 — src/lib/authorAssignments.ts

Do not surface a source list inside the public article; facts are stated in-line in Blinto’s voice.

Baseline Research

Core use case: Live chat, shared inbox and ticketing across chat, email and social, with an AI agent (Lyro) answering from merchant-supplied knowledge, plus Flows automation.
Target merchant: Small-to-mid Shopify stores wanting to deflect routine support; teams of one upward.
Current App Store rating: 4.8 (checked 2026-10-02) — highest-volume review base in the Blinto set so far.
Review count: 1,349 (86% five / 7% four / 1% three / 2% two / 4% one). Re-check before publication if cited.
Built for Shopify: Yes. Launch date: July 14, 2014.
Pricing model: Conversation-based, with separate AI and Flows allowances and paid add-ons; two public price lists that disagree.
Free plan/trial: Free plan exists; 7-day full-featured trial observed in-app.
Languages: English, German, French, Italian, Portuguese (Brazil), Spanish per the listing.

Developer Claims

Claim	Source	Verified?	Notes
“Lyro doesn’t make things up… every response is accurate and helpful”	Vendor Lyro page	Contradicted in our test	One comparative catalogue question answered incorrectly with a confident, sourced-looking reply. One instance, not a rate.
Lyro answers only from supplied knowledge and does not improvise	Vendor marketing	Not established	The wrong product exists in the catalogue, so the failure looks like faulty retrieval/comparison rather than invention; cause not determined.
67% average resolution rate	Vendor Lyro page	Not tested
Powered by Claude (Anthropic AI)	Vendor Lyro page	Not verifiable by us	Stated by vendor; not independently confirmed.
Widget installs with no code	Listing + in-app	Yes	Single theme app-embed toggle; app confirmed installation.
Email creates tickets in the Inbox	In-app confirmation	Partly	Connection confirmed; no inbound email was sent to verify ticket creation.
Chat, email, Instagram and WhatsApp in one inbox	Listing	Partly	Chat and email configured; social channels not tested.
Lyro Actions (refunds, order edits)	Listing	Not tested
Live visitor list, live typing, operating hours, multilanguage	Trial modal	Partly	Live visitor list observed; others not tested.
7-day free trial	Listing + in-app	Yes	Banner showed 7 days left throughout.

Merchant Feedback Patterns

Common positives (recent reviews, Aug–Sep 2026): intuitive setup and dashboard; Lyro deflecting common questions, with one merchant citing roughly 90% resolution; time savings for small teams; smooth Shopify integration; consolidated chat/email/social inbox; reported conversion improvements.

Common complaints (1–3-star sample, Jul 2025 – Jul 2026, checked 2026-10-02):

Cannot export customer conversations (Jul 2026, developer reply).
Onboarding support dropped off after initial meetings (Jun 2026, developer reply).
Widget positioning harmed the mobile experience (Jun 2026, developer reply).
Pricing and feature descriptions differ between the Shopify listing and Tidio’s site; bot behaviour degraded after an upgrade (May 2026, developer reply).
Charged twice with a refund refused and unresponsive billing support (May 2026, no reply).
A duplicate installation silently prevented messages arriving (May 2026, developer reply).
Site scanner non-functional and data from several websites mixed together (Feb 2026, developer reply).
Complex pricing; an upgrade needed within a day; real cost reaching ~$300/month against an advertised ~$59 (Jul 2025, developer reply).

Interpretation boundary: Sourced patterns from a small negative sample (about 7% at three stars or below), not Blinto findings. We did not reach billing, export, social reconnection or mobile testing. Notably, the pricing-discrepancy complaint matches what our own research found independently, and the “scanner mixed data from several websites” complaint is adjacent to our accuracy finding without confirming it.

Missing Evidence / Questions

Why Lyro named the wrong product: stale crawl, missing Products knowledge source, or failure to compare prices. Checking Knowledge › Data sources and Knowledge › Products would likely answer this.
Whether the myshopify URL crawled at onboarding versus the live nestly.blinto.co storefront caused a catalogue mismatch.
How Lyro performs across a representative question set (policy, stock, order status, comparatives).
Whether the Review control retrains or corrects future answers.
Whether domain verification completed.
Whether inbound email actually created a ticket.
Munna’s own impressions: setup difficulty, confidence in the AI, value judgement.
ClickUp task URL.

Editorial Evaluation

Strengths

Fast, genuinely no-code deployment with self-verified installation.
Strong merchant-side supervision: AI attribution, source labelling, per-reply review controls, inline rating events and a CSAT panel.
Graceful, accurate self-correction when challenged.
Channel-level control over where the AI operates; rich visitor context in the Inbox.

Limitations

A confident, incorrect answer to the first factual question asked, against explicit vendor marketing to the contrary.
No uncertainty signal or handoff before answering wrongly.
Two official price lists that disagree on names and prices; free Lyro allowance is lifetime, not monthly.
Domain verification up to 48 hours; “via tidio.com” until then.

Best Fit

Small Shopify stores wanting live chat immediately with AI deflection of routine questions, where a human reviews AI answers.

Poor Fit

Merchants wanting unattended AI on a catalogue where specifics matter; merchants needing conversation exports; merchants needing cost certainty from published plans.

Alternatives to Consider

Gorgias, Re:amaze, Richpanel, DelightChat, Crisp, Chatra, Zipchat (Batch 3 candidates, none reviewed by Blinto yet).

Claims We Must Not Make

Do not state an error rate or imply Lyro is frequently wrong; we tested one question.
Do not claim to know why the answer was wrong; the cause was not established.
Do not claim Lyro invents information absent from the catalogue — the product it named exists.
Do not claim Flows, social channels, Lyro Actions, exports, analytics or mobile behaviour work or fail; none were tested.
Do not claim email ticket creation was verified end to end.
Do not state support response times or billing outcomes; the trial was never completed.
Do not present sourced billing, export or duplicate-install complaints as reproduced.
Do not repeat the 67% resolution-rate or Claude-powered claims as verified.
Do not compare like-for-like against the loyalty reviews; different category and store.
Do not describe the reviewer’s subjective experience until Munna supplies it.

Blinto Review Quality Score — Internal

Quality area	Score	Maximum	Why this score
Hands-on authenticity	18	20	Full path captured from install permissions to the logged consequence of an AI failure, including both sides of the conversation; deductions for the untested cause, unverified ticket creation and no billing event.
Merchant usefulness	19	20	Answers the actual AI-buyer question, supplies a pre-launch test procedure, quantifies the lifetime free allowance and the two price lists.
Research depth	14	15	Listing, both review filters, vendor pricing page and vendor Lyro page with the exact marketing claim; the official help article was 403 and is recorded as unavailable rather than cited second-hand.
Evidence/screenshots	9	10	Eight purposeful captures including both halves of the failure and its merchant-side record; no capture of the knowledge-source configuration that would explain the cause.
Balanced criticism	10	10	Credits the recovery and the supervision tooling, states the contradiction plainly, and bounds the finding to one question on one catalogue.
Original findings	10	10	A screenshot-backed contradiction of a headline vendor claim, the full consequence trail, the policy-versus-comparative distinction, the lifetime free allowance and the pricing conflict — none available from any competitor roundup.
Decision clarity	9	10	Recommend-with-supervision verdict, concrete pre-launch test, named triggers for looking elsewhere.
Writing/editorial quality	5	5	House structure, held to the requested 12–14 minute read, consistent evidence labelling.
Total	94	100

Status: 90–100 — Strong, publishable. Scored on 2026-10-02; the article was published the same day with text identical to this record. At 94/100 it clears the 81+ auto-approval threshold in AGENTS.md §10, so manual SEO Check was not required.

Editorial Verdict Lock

Verdict locked: Yes — article published by the project owner on 2026-10-02 with text identical to this repository version.
Merchant decision: Recommend for small stores that will supervise the AI and confirm pricing in their own admin; not for unattended AI on detail-sensitive catalogues.
Best for: Small teams wanting live chat today with AI deflection under human review.
Not for: Unattended AI, conversation-export needs, or merchants needing cost certainty from published plans.
Verdict notes: Consistent with the supplied evidence and the eight published screenshots.

Founder Intelligence — Internal

What the app does exceptionally well

Deployment: six onboarding screens, one theme toggle, and the app proves its own installation to the merchant.
Supervision design: every AI reply is attributed, source-labelled and reviewable, and customer ratings surface against the AI agent by name. This is the best AI-accountability surface we have seen in any app reviewed so far.
Recovery: challenged on a wrong answer, the agent conceded and produced the correct one immediately.

Merchant friction discovered

The agent produced a confident wrong answer to a trivially checkable question, with no uncertainty signal, directly contradicting the product’s own headline marketing claim.
The entire knowledge base is established by one URL field during onboarding, which sets a merchant up to over-trust the result.
Two official price lists disagree on plan names and prices; the free Lyro allowance is a lifetime 50, which reads as monthly at a glance.
Domain verification takes up to 48 hours and sends “via tidio.com” until complete.

Growth opportunities

Align the Shopify listing and the website price lists; recent one-star reviews name this directly.
Make the lifetime nature of the free Lyro allowance explicit at the point of signup.

Product opportunities

Add an uncertainty threshold for comparative or numeric catalogue questions — answer, hedge or hand off rather than assert.
Surface which knowledge source produced an answer, not just that one did, so a merchant can fix the root cause from the Review control.
Prompt merchants to run a question set against the Playground before the agent goes public.

Questions for the founder

For catalogue comparisons, does Lyro query live product data or a crawled snapshot?
Does the thumbs-down/Review action feed back into future answers?
How is the “every response is accurate” claim validated internally?

Blinto fit

Product/UX opportunity

Evidence-based reason: Deployment and supervision are strong; the gaps are an over-confident answer path with no uncertainty handling, an onboarding flow that encourages over-trust, and inconsistent price communication — all product, UX and positioning problems rather than engineering failures.

SEO & Content Direction

Target merchant: Shopify merchants evaluating AI support agents and asking whether they can be trusted unattended; merchants comparing Tidio pricing.
Merchant problem: Knowing how far to trust an AI support agent, what supervision exists, and what it really costs.
Primary search intent: Commercial investigation.
Primary keyword/topic: Tidio review
Secondary topics/questions: Tidio pricing, is Tidio free, Lyro AI accuracy, Tidio vs Gorgias, Tidio AI chatbot Shopify, Tidio free plan limits.
SERP/content gaps: Roundups restate Tidio’s feature list and repeat the resolution-rate claim; none show an AI answer failing and being corrected, or reconcile the two price lists.
Blinto differentiated angle: A screenshot-backed AI accuracy failure and its full merchant-side consequence trail, plus the pricing-source conflict.
Internal links: First Customer Support and Helpdesk review — no same-category related links will render yet. Consider linking from future Gorgias/Re:amaze/Richpanel reviews back to this one.
CTA direction: Run your ten real questions through the agent before it goes public; confirm plan limits in your own admin; start domain verification on day one.
Required disclosures: Level 1 test of chat/AI/email setup only; development store; trial state; one question tested; Flows, social, Actions, exports and billing untested; pricing and review data checked 2026-10-02.

Thumbnail Concept and Generation Prompt

Concept: A traveller at a grand help desk in a sunlit hall. A softly glowing attendant gestures confidently toward one doorway, while the hall’s polished counter reflects a different doorway entirely — the answer given and the answer that was true, not quite matching. Reads as “confident, and not quite right.”

Prompt:

Generate a single image and export it as a PNG file, 16:9 landscape, 1792×1024 pixels or larger.

This must look like a soft, hazy, atmospheric vintage oil painting — NOT a sharp digital illustration. Softness and abstraction are the most important requirement.

Style: a sun-bleached impressionist oil painting on rough canvas, in the manner of a faded plate from an old art book. Thick, loose, visible impasto brushstrokes everywhere, including the foreground. Strong canvas texture and pigment variation across the whole surface. Low contrast, high-key light: warm cream, ivory, pale ochre, dusty gold, muted olive, soft grey-blue. No pure blacks, no deep shadows, no saturated colour. Everything slightly out of focus, as if seen through warm afternoon haze; edges lost and found, silhouettes soft, forms suggested rather than drawn.

Scene: a small lone traveller with a simple bag stands at a long polished counter in a vast sunlit hall with tall windows. Behind the counter, a softly glowing pale figure gestures confidently toward one of several distant archways. In the counter’s reflection, the gesture appears to point toward a different archway instead. The two do not match. The traveller is the only soft focal point and is small in the frame; the hall, counter and archways dissolve into broad blurred masses of colour.

Light: high in the upper background, through a tall window, a single soft pastel blue ring shape glows faintly like a distant sun seen through haze — feathered edges, varied opacity, painted with the same loose brushwork, never a logo or a clean circle.

Depth: foreground moderately soft, middle ground much softer, background almost abstract colour fields.

Strict rules: no text, no letters, no numbers, no logos, no signage, no UI elements, no screens, no phones, no speech bubbles, no watermark. No photorealism, no 3D render, no glossy digital painting, no fine detail, no sharp edges anywhere, no faces with detail. Wide composition with generous empty space left and right so the image crops safely.

Output format: PNG.

Source colour note: The ring’s tint derives from Tidio’s blue brand mark and the blue chat launcher, softened into a pastel against the painting’s cream light. Verify the exact brand hex from the live listing before generating; do not let the image tool reproduce a crisp logo.

Public Article Authenticity Gate

Publication Gate

The internal research preview will be /research/tidio/. The SEO Check handoff must use the rendered review URL /reviews/tidio/, never the research URL.

Tidio: AI Chatbot & Live Chat (theme embed: "Tidio Chat Widget"; AI agent: "Lyro")

Blinto Review Quality Score