Skip to main content
14 min readGetting Started

Troubleshooting Common Setup Errors

A symptom-to-fix reference for the errors agencies actually hit during setup — exact messages, what each one really means, and the fastest way to resolve it.

Most setup problems produce a specific message or a specific silence, and both point at a real cause. This guide lists the errors agencies most often hit — quoted exactly as they appear on screen — with the underlying cause and the fix, so you can match what you are seeing and resolve it without guesswork.

Each section covers one area of the platform. Scan for the message you are seeing, or use the silent-failure sections (auto-respond, email delivery) when something simply stops working without an error.

The messages are quoted verbatim on purpose: if the text on your screen matches a row here, the cause and fix in that row apply — no interpretation needed.

Payments and account lock (your agency subscription)

These errors relate to your own agency subscription, not your customers' billing. When your subscription or trial lapses, the platform locks navigation but deliberately keeps billing reachable so you can always fix it yourself.

SymptomCauseFix
`Navigation Locked` overlay with `Your account is locked due to non-payment...`Your agency has no active subscription or trialClick `Restore Access` — it opens the payment modal. While locked, `Resources -> Account` stays clickable, so billing is always fixable
`Your card was declined...` (or a specific CVC / insufficient funds message)Your bank rejected the charge — the message maps to the actual decline reasonFix the stated problem (correct the CVC, add funds, or use a different card) and retry
`Additional authentication is required. Please complete the authorization.`Your bank requires 3D Secure confirmationComplete the authorization in the popup your bank presents. If you dismiss it, the payment stays incomplete even though it looked submitted. You get up to 3 retries per session
`Action Required: Payment Overdue` bannerA renewal payment failed on your saved payment methodUpdate the payment method via `Manage Subscription`, which opens the Stripe billing portal
Worth knowing

The 3DS case is the sneakiest: the payment form looked submitted, but until you complete your bank's authorization prompt in the popup, the payment stays incomplete. If access does not restore after paying, check whether a 3DS popup was dismissed and retry — you get up to 3 retries per session.

Sales Intelligence search errors (DataForSEO)

Sales Intelligence searches run on your own DataForSEO credential. When a search fails with a mapped message, the message itself tells you what to do. When it fails with the generic message, the failure was unmapped or internal — check the credential's last error in Sales Intelligence settings for the real reason.

Credentials are deliberately never disconnected on failures, so a failing key stays in place for you to inspect and fix rather than silently vanishing.

SymptomCauseFix
`An error occurred while searching. Please try again.`An unmapped or internal failureCheck the credential's last error in Sales Intelligence settings — that shows the underlying reason
`Your search credits have run out.`The DataForSEO account balance is emptyTop up credits on dataforseo.com
`Your server's IP address needs to be authorized.`The DataForSEO account restricts API access by IPWhitelist the server IP in your DataForSEO account settings
`You've reached your daily search limit.`The daily search cap has been hitWait for the limit to reset, or raise the limit
`The location data needs to be refreshed.`The location string could not be resolvedRetry with a 'Dallas, TX' style location format
`Test Connection` passes but real searches failThe connection test only checks authentication, not the paid search endpoints — commonly a brand-new DataForSEO account that has not finished DataForSEO's own activation stepsFinish account setup on dataforseo.com and add credits, then search again
Search sits pending, then shows `Search timed out...`The search returned nothing, so it shows as still-pending until it times outCheck the credential's last error and the DataForSEO account, then retry
A prospect on your public lead-capture form reports `We could not search for businesses right now. Please try again in a moment.`Any underlying search failure — the public form deliberately never shows prospects the real errorDiagnose from your own account: run the same search there and read the actual error
Worth knowing

A low DataForSEO balance shows passively in the workspace header — there is no alert email, so make a habit of watching it.

OpenRouter and AI provider errors

AI features (auto-respond, AI reply suggestions, AI Insights) run on the AI provider configured at `Platform -> AI & Automation -> Artificial Intelligence`. The failures below are the ones agencies actually hit — and the most dangerous one is silent: live auto-respond shows no banner when OpenRouter fails.

SymptomCauseFix
`The AI model configured for your AI provider is no longer available on OpenRouter...`The configured model was retired or renamed on OpenRouter (e.g. old Claude 3.5 ids)Edit the provider at `Platform -> AI & Automation -> Artificial Intelligence` and pick a current model — the wizard now blocks retired ids
`OpenRouter rejected the saved API key...`The key was pasted with a `Bearer ` prefix or hidden whitespaceRe-paste just the key itself (it starts `sk-or-`)
`Your OpenRouter account has insufficient credits...`The OpenRouter balance is emptyTop up at openrouter.ai
Auto-respond replies simply stop, with no banner or error anywhereLive auto-respond has no banner for OpenRouter credit or model failures — replies silently stopCheck the rule's `Activity` log for error rows, then fix the provider (credits or model)
`No AI provider with a valid API key found. Please configure your AI provider in Settings.`No AI provider is configured, ready, and set as defaultConfigure a provider and make sure it completes setup and is the default
Worth knowing

Failed AI Insights attempts still consume the daily analysis cap, so fix the provider before retrying analyses.

Email provider errors

The single most important rule: a provider only sends once its setup wizard completes to ready. A half-finished wizard silently falls back to a non-working default — which is the usual cause of customers reporting that password reset emails return an error, and of campaign emails never arriving. Always finish the wizard until the provider shows ready, then set it as the default.

SymptomCauseFix
Microsoft 365 error mentioning `5.7.139` or `SmtpClientAuthentication is disabled`The M365 tenant has Authenticated SMTP disabled for the mailbox — an app password cannot fix thisThe M365 admin must enable it: `Set-CASMailbox -Identity "user@domain" -SmtpClientAuthenticationDisabled $false`
Gmail: `Authentication failed. Make sure you're using a Gmail App Password...`The regular account password was used instead of an App PasswordGenerate a Gmail App Password and use that — never the account password
Brevo or SendGrid: `Invalid API key. Please check your key and try again.` when the key IS correctAn IP allow-list on the provider account is blocking EMR's sending serverAdd EMR's sending server IP `49.13.0.216` to the provider's allow-list
SendGrid key verifies but sending fails with a restricted keyThe restricted key is missing required permissionsGrant `Mail Send`, `Mail Settings` and `Tracking` at Full Access
Postmark: `Failed to create domain`The domain conflicts on Postmark's side — often already registered at the rootTry a subdomain, or resolve the conflict inside Postmark
Customers report `password reset emails return an error`, or campaign emails never arrive, with no provider error shownThe email provider wizard was never completed to ready, so sending silently fell back to a non-working defaultRe-open the wizard, finish every step until the provider shows ready, and set it as the default
Worth knowing

When email "just doesn't work" with no visible error, check provider readiness first — a provider that shows anything other than ready is not the one doing the sending.

Google and Facebook review source errors

Source cards carry status badges that each mean something specific — read the badge before anything else, because it usually tells you exactly which of the fixes below applies.

BadgeMeaningFix
`Auth needed`The token expired or was revokedClick `Re-authenticate`
`Action needed`A permission problem on the connected accountFix the permission (see the rows below), then reconnect
`Limit reached`The plan's Review Source Limit is hitRaise the plan's Review Source Limit
`Upgrade`The source is not included on the current planUpgrade the plan or use a plan that includes it
`Setup needed`The connection was started but never finishedOpen the source and complete the connection flow
`Coming soon`The source is not yet supportedWait — no action available yet

Google connection and permission errors

  • Re-auth behavior differs by platform: Google always returns you to the location picker — your data is kept, just re-confirm the profile
  • Facebook re-auth self-heals and re-syncs automatically when it can, so a Facebook `Auth needed` often resolves itself after re-authentication
SymptomCauseFix
`To connect Google you must allow the "See, edit, create and delete your Business Profile" permission. Please try again and leave that box ticked.`The Business Profile checkbox was unticked during Google consent — a token refresh can never fix this, because refreshes preserve the original scopesReconnect and leave the Business Profile permission box ticked
Reviews banner: `Google isn't sharing reviews for this profile yet: ... needs Owner or Manager access`The connected Google account lacks Owner or Manager access on the listingFix the role at business.google.com, then reconnect
Insights banner: `We're unable to retrieve performance insights...`Same cause — the connected account lacks Owner or Manager on the listingFix the role at business.google.com and reconnect; insights appear within ~24h
Auto Respond never fires on a source, with no errorAuto Respond only works on Grant Access connections (`api` connected) — Public Access sources never auto-respond, silentlyReconnect the source with Grant Access

White-label and custom domain errors

Domain problems almost always come down to DNS records that were left behind. The platform expects exactly one A record for the hostname — anything else on the same name, especially an AAAA record, is the classic reason a domain never verifies or never gets its certificate.

SymptomCauseFix
The root domain redirects to the login pageNo website page is published at the `/` slugRestore a homepage in the Website builder
Custom domain never verifies or never gets SSLLeftover DNS records — a stray AAAA record alongside the A record is the classic failurePoint one A record to `178.105.206.57` and DELETE any other A/AAAA records for the hostname; Cloudflare must be DNS-only during setup
Domain verified but no SSL yetSSL issues automatically on the first HTTPS visit after verificationVisit the domain over HTTPS once; if it still fails, check for leftover AAAA records
Blank page after connecting Google or FacebookNo primary domain is set on your account, so the callback has nowhere to landSet a primary domain under `Platform -> White-Label -> Domains`, then reconnect
Favicon upload appears to do nothingThe file was rejected — it must be `.ico` or `.png` and under 1MB; the preview clears on rejectionCheck the file type and size, then upload again
Worth knowing

If the domain sits behind Cloudflare, keep the record DNS-only (grey cloud) during setup — proxying it before verification completes is a common reason the domain never verifies.

AI Insights errors

AI Insights runs share the same AI provider as auto-respond, so a broken provider fails analyses too — and because failed attempts count toward the daily cap, a broken provider can burn a whole day's attempts in minutes.

SymptomCauseFix
`Daily attempt limit reached (3/3). Please try again tomorrow.`A configured cap of 3 analyses per organization per day — failed and in-progress attempts count toward itFix the AI provider BEFORE retrying, so failures do not burn the cap
`Not Enough Reviews`The selected period has fewer than 5 text reviewsWiden the period or wait until at least 5 text reviews exist
`Analysis timed out...`A stuck run auto-fails after ~10 minutesRetry once the AI provider is healthy

Auto-respond silent skips

When 'auto-respond does nothing', there is usually no error at all — the review was skipped by design. Check these conditions in order before suspecting the AI provider.

SymptomCauseFix
No reply sent, no error shownThe review is older than 2 daysExpected — auto-respond only handles fresh reviews; reply manually or use a bulk `Old Reviews` run
No reply sent, no error shownThe review already has a replyNothing to fix — already-replied reviews are always skipped
No reply sent, no error shownThe review source is not Google or FacebookExpected — only Google and Facebook support auto-respond
No reply sent, no error shownThe connection is Public Access instead of Grant AccessReconnect the source with Grant Access
No reply sent, no error shownThe customer's plan lacks Auto-RespondingAdd Auto-Responding to the customer's plan
Worth knowing

Bulk `Old Reviews` runs are different: they record per-review outcomes in the progress modal (e.g. `Review already had a reply`, `Connection can't post replies`, `Plan doesn't include auto-respond`, `AI couldn't write a reply`). The bulk button only exists on rules scoped to a specific organization.

Login and access errors

Two access errors come up repeatedly during team and customer onboarding, and both look worse than they are.

SymptomCauseFix
`419 | Page Expired`The page sat open too long and the session token expiredRefresh the page and retry the action. If it recurs on a custom domain, that suggests a cookie/domain mismatch worth reporting
Invite link shows `The email has already been taken.`The person already has an account and used the register link from the invitationThe invitation email also contains an accept link for existing accounts — use that one; it routes existing users to sign-in and joins them
Worth knowing

A one-off `419 | Page Expired` is normal and harmless — a stale tab, not a bug. It only becomes worth reporting when it keeps happening on a custom domain.

Common questions