Something not right? Start here.
A focused list of the issues we see most often, what causes them, and what to try before writing to us. The closing block has our address — and we would rather hear about it than have you guess.
Three checks that resolve most reports.
- Check status.meetindi.com — platform-level incidents are posted there before email.
- Confirm forwarding — your business line should forward to your MeetIndi number, not an old test number from setup.
- Check role — some surfaces are role-scoped. If a page is missing, your account role may not include it.
Indi answers, then drops the call within seconds.
- Caller hears the greeting, then silence
- Call disconnects under five seconds
- Dashboard shows the call but the transcript is empty
A handoff between Twilio and the voice runtime did not establish — usually a forwarding mis-set, an auth refusal at the model layer, or a single-carrier path that is misbehaving on that leg. It is almost never a database issue.
- Check status.meetindi.com first — if voice is degraded, no action is needed on your side
- Confirm your business line is forwarding to your MeetIndi number, not an old test number
- Place a test call from a different carrier or device to rule out a single-path fault
- Note the call SID from the dashboard call detail before reporting — it lets us trace the exact session
Three or more failed test calls in a single hour, or the dashboard never registers the call at all.
Indi sounds robotic, slow, or stalls mid-sentence.
- Long pauses between turns
- Clipped or stuttering audio on the caller side
- Indi repeats a question or restarts a sentence
Almost always transient — model latency on the provider side, network jitter on the call leg, or an unusually long caller turn the model is still transcribing. Sustained patterns are a different signal.
- Single-call symptoms usually self-correct on the next turn
- Persistent across multiple calls in the same hour: check status.meetindi.com for a voice-runtime advisory
- For a known noisy line, ask the caller to move to a quieter spot — model audio benefits from clean input
- Capture the call SID and approximate timestamp before writing to us
Five or more calls with the same symptom inside a single business day, especially clustered to one hour or one caller.
A call happened but it is not in my Action Queue.
- Missed call on your business phone, nothing in the dashboard 60 seconds later
- Call appears under Calls but never surfaces in Action Queue
- Caller says they spoke with Indi but no record exists
Three common reasons: the post-call pipeline is still running, the call landed on a surface your role does not include, or the call was sub-30-seconds and was not classified into the queue.
- Wait 90 seconds and refresh — the post-call pipeline runs after hangup and is intentionally async
- Action Queue only shows unactioned items; check the Calls page for the full list
- Confirm your role: agents see assigned leads only, managers see their team, owners see everything
- Sub-30-second calls are kept in Calls but not in the queue — that is a feature, not a gap
A call that completed normally (over 30 seconds, transcript visible) and is still absent from Calls five minutes after hangup.
Lead notes are not pushing into FUB.
- Call completed in the dashboard but the FUB lead has no MeetIndi note
- Note appears in FUB with placeholder text or missing fields
- A "Sync deferred" flag on the call entry
Most often a per-tenant API key that was revoked or expired on the FUB side. Less often a FUB rate-limit hit during a burst, or a person-ID mapping that has drifted since the last reconnect.
- Settings → Integrations → FUB: confirm the status reads "Connected", not "Reconnect needed"
- Reconnect needed: re-paste the API key in the masked input — the existing key is never returned in plaintext
- For a specific missing note, open the call detail and use "Resync to CRM"
- Watch the Calls page for entries flagged "Sync deferred" — they retry automatically and should clear within an hour
A reconnect does not clear the warning, or a deferred sync still has not caught up overnight.
"Consent missing" — outbound SMS will not send.
- Clicking a follow-up SMS action returns a "consent required" message
- Lead clearly expected a text but the send is blocked
- No SMS history with this number even though calls happened
By design. CASL is enforced structurally — every outbound SMS requires an immutable sms_consent_log entry. Without one, the code path refuses. There is no admin override.
- Indi captures verbal SMS consent during the call; if the caller declined or the prompt did not run, no consent exists
- Re-engage by phone first; on the next call, the caller can verbally opt in
- Inbound SMS from the lead creates implied consent for a reply window — let them text first
- The consent log is append-only; revoked consents are flagged but never deleted
A lead who you can prove opted in (call SID with the consent moment) but is still flagged "consent missing" — that is a logging bug, not policy.
Bookings are not syncing both ways.
- Caller says they booked, the dashboard shows nothing
- Dashboard shows "Booked" but the calendar event is missing
- Bookings land outside working hours and the agent never sees them
Per-agent calendar tokens expire periodically. Out-of-hours bookings reroute to the next available agent silently — that is intentional. Double-book detection refuses the slot quietly when the agent is already busy.
- Settings → Integrations → Calendar: each agent reconnects independently from this surface
- Indi only says "confirmed" after both caller-side acceptance and calendar acceptance — check the transcript for the exact wording
- Dashboard "Booked" requires both halves; if one fails, the lead lands as "Qualified" instead
- Working-hours config is per-agent; out-of-hours bookings reroute and the original agent will not see them
A confirmed booking visible on the calendar that never reached the dashboard.
Cannot log in, or the session expires within minutes.
- Login loop — credentials accepted, immediately bounced back
- "Session expired" within minutes of a fresh sign-in
- Account exists but the password is rejected
Token model is 15-minute access / 30-day refresh. A revoked refresh token forces a full re-auth. Browsers blocking third-party cookies on the auth domain produce a similar symptom for a different reason.
- Sign out fully and back in to force a refresh-token re-issue
- Try a private window or a different browser to rule out cookie-domain blocks
- "Account not found" almost always means a typo in the email — pilot tenants are scoped, so a different domain lands on a different tenant
- Some screens are role-scoped — if a specific page rejects you, your role may not include it
Repeatedly ejected within minutes across two browsers, or a confirmed-correct email that says "Account not found".
My usage or invoice does not match what I expect.
- Minute count looks higher or lower than your call log suggests
- Proration on a plan change does not match the day count
- Expected refund has not posted
Three rules drive most of the discrepancies: minutes round up per call (not per month), plan changes prorate by day, and refunds post the next billing cycle rather than immediately.
- Owner role only — billing surfaces are owner-scoped; manager, ISA, and agent roles do not see this
- Settings → Billing → Usage detail shows per-call line items with the rounding visible
- Plan-change entries show the proration window in days; multiply by daily rate to verify
- For a refund question, the public refund-policy page is the rule of record
A line-item total that still does not reconcile after reviewing the per-call detail.
The recording or transcript is missing.
- "Recording unavailable" on a call detail page
- Audio is missing but the structured fields exist
- Transcript is empty even though the call lasted longer than 30 seconds
Three retention rules: sub-30-second calls are not retained at all, calls older than your DPA window have rolled off as scheduled, and the transcript pipeline can lag the audio by up to a minute on busy days.
- Sub-30-second calls: not retained. The caller hung up before the interaction produced enough usable call content; no audio or transcript is kept.
- Older than your retention window: rolled off as scheduled. The call event audit log still exists.
- Within retention but missing: open the call detail and try "Reprocess" — it surfaces when our pipeline previously failed
- Fresh call with no transcript: refresh after 60 seconds; the transcription stage runs after the audio is finalized
A recent call inside your retention window where "Reprocess" does not recover the audio or transcript.
Still stuck? Tell us what you see.
Include the call SID, the approximate time, and the symptom as the caller experienced it. We will come back with specifics — and we will tell you honestly if the cause is on our side, on yours, or somewhere between.
Related: Security, Privacy Policy, CASL compliance.