AI that reads your deeds, qualifies your leads, and runs your property desk

One flat subscription covers all eleven modules, every dashboard and every property record — Qelvantiq.AI is not sold per seat, per listing, per parsed document or per lead. Hire three more negotiators, open a second branch, take on another four hundred managed units: the monthly figure does not move. The only other money on the table is the one-time onboarding fee above, and optional add-on hardware if you want it. There is no tier to decode, and nothing that quietly meters you in a busy month.

Because a property desk is not live until your book is inside it. That fee covers the work of getting there: importing your existing records, vendors, landlords and tenants, loading live listings with photos, measurements, tenure and EPC bands, wiring portal and website enquiries into one deduplicated queue, and tuning routing rules, contractor SLAs and valuation presets to how your teams actually run. It starts at $1,000+ for a single desk, $2,500+ where historical deals and integrations need migrating, and $5,000+ for multi-branch or full-archive work. Paid once, before your first monthly cycle — not again every year.

Neither, in the way you are worried about. Nobody is asking you to rip out Yardi, RealPage, Entrata or MRI, or to give up a portal feed that took a year to get approved — Qelvantiq.AI reads from them. What it replaces is the layer of human work wedged between them: the same tenancy re-keyed into three systems, listing copy pasted into four portals, the spreadsheet of break dates one person maintains and nobody else can find. One honest caveat — Rightmove and Zoopla publish partner feeds, and Yardi, RealPage, Entrata and MRI issue credentials per client, so those need an agreement in place. We tell you which of yours are gated on the demo call, not after you have signed.

Yes, and most desks should. The subscription already covers the whole platform, so the only decision is where onboarding starts — usually the module attached to whatever is bleeding: lead response for a sales team, document extraction for a compliance lead, work orders for a property manager drowning in tickets. Bringing the rest online later is a configuration change your admin makes, not a new quote and not a second migration.

It gets caught before it reaches the record. The Document Extraction Module scores every field it pulls — contract value, parties, dates, break clauses, covenants — and anything it is unsure of goes to an exception queue on the Central Dashboard instead of being written silently. A conflicting rent schedule, an unreadable exhibit, a missing signature block: flagged, not guessed. Every value keeps a link back to the page and line it came from, so your reviewer opens the source beside the number rather than taking it on faith. Your team signs off. The AI never gets the last word on a legal document.

The controls sit inside the modules doing the writing, which is the only place they are worth anything. Marketing Content Generation drafts against a screened vocabulary, so descriptions stay about the property. The Lead Conversion Assistant scores on response behaviour, budget fit and service intent — never on a protected characteristic. Outbound contact from the Client Communication Co-Pilot and the Customer Chat Bot honours consent state, quiet hours and opt-outs at the point of send. And every generated description and routed enquiry keeps an audit trail of what was produced, by which module, and when — which is the artefact that actually shortens audit week. Qelvantiq.AI is built around Fair Housing, TCPA, ESIGN and CCPA rather than retrofitted to them.

Your documents and property records stay in your own tenant, scoped by role — a maintenance coordinator sees work orders without seeing valuation files. They serve your workspace and nothing else: not training a shared model, not enriching another agency's data. If that is still too close for your compliance officer, the same platform deploys inside your own cloud, VPC or on-prem, and the subscription is unchanged — only onboarding differs. Export and deletion run through your workspace owner on request, and records leave in a structured form you can load somewhere else.

Because somebody is. Every agent here is run as a production system by Nestack Agent Care, our sister platform: 190 catalogued failure modes for real estate, 82 of them rated SEV-1, 965+ baseline evaluation cases, and 24/7 monitoring with telemetry, guardrails and human review. Drift is the normal failure mode of AI software, and it shows up as quietly worse output rather than an outage — which is precisely why it is monitored instead of trusted. If an agent starts wandering at three in the morning, a crew sees it before your landlords do.

Concretely: the Saturday-evening enquiry is answered the moment it lands and arrives scored on Monday instead of cold; renewal and break dates come out of the actual lease rather than out of whoever remembered to diarise them; a leak reported at eleven at night is triaged, matched to an approved contractor and dispatched without a phone call; listing copy is written once and pushed everywhere. Desks built on this aim at roughly 3x faster enquiry response, around 70% faster work-order turnaround, and contract audit time down by about 85%. Those are targets drawn from the workflows, not a guarantee bolted to your name — the post-launch review measures yours.

Fair scepticism, and the answer is further up this page rather than in this paragraph — eleven working dashboards, each at a live URL, none of them a roadmap slide. The difference that matters is that these modules act rather than suggest. The chat bot qualifies the visitor, writes to their CRM record and queues the next action. Maintenance Operations assigns the contractor. The Client Communication Co-Pilot sends the chase for the missing document. A person is pulled in when the case genuinely needs one, which is a design decision rather than a limitation. A chatbot answers a question; this runs the desk between the questions.

Put the monthly figure next to what it displaces. It replaces ten-plus point tools with one property record, and the arithmetic usually ends the conversation: $200 a month is under a day of an administrator's time, and a single instruction lost because an enquiry sat in an inbox over a weekend costs more than the year. One avoided compliance penalty covers it outright. The Central Dashboard also answers the second question your owner will ask — parse volumes, field accuracy, risk flags and exception queues on one screen, so portfolio risk becomes something they can see rather than something they have to ask you about.

Onboarding is run for you, not handed over as a login and a help centre. Plus puts a dedicated onboarding specialist across the rollout; Enterprise gives you a named onboarding manager, a committed go-live date and a post-launch review. Training is delivered live and recorded, so a new starter in March does not need you to repeat January. After go-live, Agent Care watches the agents around the clock, and your workspace owner has a route to a human for the things monitoring cannot fix on its own.

Forty-five minutes on a video call, free, in three steps: your details, the modules you want to see, and a slot that suits you. Then we open the real dashboards — not a deck — and work through how your records would map onto them. Bring one messy lease and one honest week of enquiries. Those two things tell us more about whether Qelvantiq.AI fits your desk than anything we could put on a slide, and they tell you whether the extraction and the scoring hold up on your own paperwork.