FAQ

Quick answers to common questions.

If your question isn't answered here, contact us — we'll add it to the list and reply same day.

Pricing

How we charge

How is OmniTakeoff priced?: Per-estimator seat pricing with three tiers: Starter ($199/seat/mo), Professional ($399/seat/mo), and Enterprise (custom). All tiers include unlimited bids and projects. Volume discounts kick in at 10+ seats.
Is there a free trial?: We offer four-week pilots — paid but credit-back if you don't see measurable outcomes. The pilot includes onboarding sessions, dedicated success engineering, and outcome-target commitments up front.
Can I add or remove seats mid-cycle?: Yes. Seat changes are pro-rated. Add seats: charged immediately for the remaining cycle. Remove seats: credited on the next invoice.
Do you charge per bid or per document?: No. Unlimited bids and documents on every plan. AI usage is included; we don't pass through model token costs separately.

Accuracy

How accurate is the AI takeoff?: Out of the box, recognizer accuracy is meaningfully lower than where it ends up after a few weeks of customer-specific active learning. Recognizers improve substantially when customers run the confirm/edit/reject review queue on their own drawings; specific accuracy targets depend on trade and drawing convention. Honest answer: there isn't a single number that applies to every customer.
What happens when the AI is wrong?: Low-confidence items land in the review queue. The estimator confirms, edits, or rejects each one. Every decision becomes ground truth that retrains the recognizer for the next bid.
Do I need to train it for my drawings?: Not for the first project — the shipped recognizer covers most common symbols. For trade-specific or proprietary symbol conventions, the active-learning loop accelerates accuracy from project 2 onward without manual training data labeling.
Can the AI hallucinate items that aren't there?: Yes, generically. We mitigate via three signals: vector-text grounding (does the PDF text near the bbox match?), symbol-library cross-check (does the org have an active definition?), and trade-context (does the claim's CSI match the sheet's classified trade?). All three feed a confidence-adjusted score that gates auto-OK.

Security & privacy

Where is my data stored?: AWS US-East and US-West regions by default. EU regions available on Enterprise. Customer data is encrypted at rest (AES-256) and in transit (TLS 1.3). Postgres Row-Level Security enforces tenant isolation at the database layer.
Are you SOC 2 compliant?: SOC 2 program in progress with an independent audit firm. Current attestation status is shared with prospects under NDA via the Trust Center; we don't publicly claim certifications we have not yet earned.
Do you train AI models on customer data?: Per-org models train only on that org's confirmed annotations. We never train cross-org models on customer data. Anonymized aggregate metrics may inform product priorities but never the recognizer.
Can I delete my data?: Yes. Production GDPR + CCPA data-subject rights endpoints export or hard-delete an org's data within 30 days of request. Settings → Privacy → Request data export / deletion.

Integrations

What does OmniTakeoff integrate with?: QuickBooks Online (live), Sage 300 (in progress), Procore (live), Autodesk Build (in progress), Bluebeam, Revit (IFC import), Slack, Microsoft Teams, Dropbox, Box, Google Drive, OneDrive, SharePoint, RSMeans. Full directory at /integrations.
Do you have an API?: Yes. Full REST API with OpenAPI 3.1 spec. TypeScript and Python SDKs generated from the spec. HMAC-signed webhooks for lifecycle events. See /api for endpoints + the developer hub.
Can you build a custom integration for us?: Enterprise tier includes dedicated engineering hours per quarter for custom integrations. For commonly-requested integrations (e.g. ERP sync, custom CRM), tell us — when 2+ customers ask we typically prioritize on the roadmap.

Pilots

How does a pilot work?: A scoped, paid window using OmniTakeoff on real bids (typical scope is 4–8 weeks across 2–3 bids; exact length agreed in the pilot scoping call). We commit to a specific outcome target up front (e.g. an estimator-time-saved or bid-throughput goal calibrated to your baseline). At pilot-end, we measure together. If outcomes hit, you sign a paid plan. If not, the platform-license cost is refunded. See /pilot-faq for the full mechanics.
What's expected of us during the pilot?: Three things: upload at least 3 real plan sets, walk the review queue daily, and complete the weekly check-in (30 min). The check-ins are how we calibrate the recognizer to your trade and surface workflow blockers fast.
How long is a typical sales cycle?: From first call to signed pilot: typically 1-3 weeks. Pilot length is set during the scoping call (4-8 weeks is the common range across 2-3 bids). Enterprise procurement timelines vary substantially by customer's process — large GCs with formal vendor review take meaningfully longer than mid-market customers, and we won't publish a specific range because the right number depends on your specific procurement cadence.

Question we missed?

Email questions@omnitakeoff.com or use the chat bubble in the bottom right of any page.