Why this order matters
A human support agent is usually the slowest way to solve a TenderB issue — not because we don't want to help, but because most questions have a faster answer in Fin (our AI assistant) or with a 20-second self-check. Escalating before you've tried those means waiting for something you could have fixed yourself.
The rule is simple: Fin first, quick checks second, human last.
Step 1 — Ask Fin in the in-app chat
Click the blue speech-bubble in the bottom-right of TenderB. Describe the issue in plain language. Fin is trained on TenderB's product docs and the patterns our team sees most often. If Fin solves it, you're done in under a minute.
Step 2 — Run these quick checks yourself
Before escalating, take 30 seconds to rule out the common culprits. Most tickets we receive are one of these:
Refresh the page (⌘/Ctrl + R) — fixes most stuck UI.
Try an incognito window — rules out extensions, cache, stale cookies.
Confirm the file is in the project's Files panel, not just the chat.
Switch AI model once if an edit didn't land on the Canvas (article: Chat changes not appearing in the Canvas).
Check you're on the right URL — standard users log in at
app.tenderb.com. Advanced and Enterprise users have their own subdomain:<your-company>.tenderb.nl(e.g.acme.tenderb.nl). Using the wrong one is a very common cause of login loops and "I don't see my projects" reports — don't use an old bookmark.Check the status page — status.tenderb.nl shows live platform availability.
When TenderB shows you an error — try the built-in fix first
Most TenderB error messages tell you exactly what to do next. Read the message first — then use the patterns below before escalating.
File upload errors
Re-upload the file. Most upload errors are transient — a second attempt usually works.
Check the file extension is on the allowed list (PDF, DOCX, XLSX, PPTX, TXT, MD, CSV, common image formats).
.zip,.rar, and unusual formats will be refused.Check the file size is within the limit. Huge PDFs with scanned images are a frequent cause — compress or split the file and try again.
Chat / AI errors
When the AI returns an error or a bad answer, don't keep pushing the same broken thread forward. Treat it like bypass surgery: route around the broken part, don't try to push through it.
Branch from the last good answer. Scroll up to the last reply that worked, click it, and continue from there. The broken turn gets bypassed and the conversation keeps its context.
Check the context window. Very long chats run out of room and start erroring out. Ask the AI to summarise / compact the conversation so far, then continue from the summary — or start a fresh chat anchored to the project's Files.
Rewrite the prompt instead of replying again. Click your message in the blue bubble, hit the ✏️ pencil icon, and edit it. The thread re-runs from that point — cleaner than a follow-up "I meant…".
Switch model. If one model keeps failing on the same step, switch to another and retry. Different models fail in different places.
If none of that moves the needle, screenshot the error (including the message text) and take it to step 3.
Step 3 — Escalate to a human (when it's genuinely needed)
Escalate if:
Fin's answer was wrong or missing, and the quick checks and built-in fixes above didn't help.
Something is clearly broken on our side (error screen, login loop, blank Canvas after multiple refreshes, status page incident).
You have a billing, contract, security, or data-deletion question — these always go to a human.
In the chat, type "Talk to a human" or "Escalate" — Fin will hand the conversation to our team.
Make the human faster — include all of this
Paste these five things into your first message. Skipping any of them is what most commonly stretches a short ticket into a multi-day one:
Workspace / organisation name
Project name or URL where the issue happens
Steps to reproduce — what you did → what you expected → what actually happened
Screenshot or short screen recording
Browser + OS (e.g. Chrome 126 / macOS 14)
Support hours
Mon–Fri, 12:00–17:00 CET, excluding public holidays. Outside that window the SLA clock starts when we next open.
Priority levels and contractual SLA (individual contracts might deviate from the levels listed below)
These are the contractual response and resolution times from Chapter 9 of the LemonB Agreement. "Response" = first human reply. "Resolution" = fix or workaround delivered. All hours count against business hours.
Priority | What it looks like | Response | Resolution |
P1 — Critical | System inaccessible or critical failure. Nobody can log in; Canvas won't load; data visibly lost. | Within 2 hours | 8 business hours |
P2 — High | Significant impact on business operations. A core workflow broken for one or more users with no good workaround. | Within 4 hours | 16 business hours |
P3 — Medium | Limited impact on business operations. Minor glitches, inconsistent behaviour, partial workarounds available. | Within 8 hours | 40 business hours |
P4 — Low | Minimal impact, general questions, feature requests, documentation gaps. | Within 16 hours | 80 business hours |
Pick your priority honestly. A tender deadline later today, an unclear AI answer, or a single file that won't upload on one browser is rarely P1. Choose the lowest priority that honestly describes your situation — it keeps the queue fair and gets the truly critical issues the response speed they need.
Use P1 only if all of these are true
P1 is reserved for widespread production incidents. Before escalating as P1, confirm every item below:
Production outage or data loss affecting most users in your workspace.
Core functionality is unavailable or compromised — nobody can log in, Canvas won't load at all, AI is completely non-responsive across all projects, or billing/checkout is entirely broken.
No temporary workaround exists (different project, Word export, different browser — none of those get users moving).
Status page checked. You've looked at status.tenderb.nl and the relevant third-party pages (see below).
Do not use P1 for
Slower page loads or general performance dips.
Issues on a trial/demo/sandbox workspace.
Glitches on a single rare device, browser or screen size.
Issues confined to a test or staging project.
A temporary workaround exists that keeps users productive for a short time.
Check third-party status before escalating
A lot of "TenderB is broken" incidents turn out to be an upstream provider. Check these before raising P1:
status.tenderb.nl — TenderB platform
health.aws.amazon.com — AWS (we run in
eu-central-1, Frankfurt)cloudflarestatus.com — Cloudflare (CDN, DNS, WAF)
If the root cause is upstream, tell us in the escalation — we'll still help coordinate, but the fix timeline sits with the provider.
Platform availability
TenderB is contractually committed to 99.8% monthly availability. Planned maintenance runs between 22:00–06:00 CET and is announced at least 48 hours in advance on the status page.
For anything else
Use the in-app chat. There is no separate support email address or phone line — every message lands in the same queue, is categorised by priority, and is handled in order.