Impact studies: measured results from systems in production
What Nezam AI is running inside client operations right now, across property management, brokerage, hospitality, and trades. Every figure below is measured, and the window it was measured over is stated. Clients are anonymized where they asked to be.
Study 01: Zero missed calls and a quarter of all new leads
Sector: property management and brokerage. Read it: /work/voice-agents-property-management
A Toronto property management company and its brokerage arm ran on inbound calls, and outside office hours the whole operation answered with a voicemail box. Two AI voice agents now answer every inbound call, capture the caller and their intent, and either resolve the call, flag it urgent to staff, or write it into the CRM as a lead.
Measured results:
- 0 calls missed. Every inbound call in the sample was answered on the first ring.
- 26% of all new leads. Of 130 new leads recorded between 8 and 22 July 2026, 34 arrived through the voice agent. Second largest source in the business, ahead of the next two listing portals combined.
- 27% after hours or weekend. 74 of 277 calls came in outside office hours, the volume a voicemail box was absorbing.
- 7 in 10 brokerage calls leave a routable lead. 120 of 167 ended with a callback number and enough context to act on.
Source and limits: call figures come from live call logs deduplicated by call ID, 167 brokerage calls over 17 days and 110 property calls over 10 days, a run rate of roughly 600 a month at current volume rather than an audited annual total. The lead export shows where leads came from, not what lead volume was beforehand, so this is a share of source and not a lift claim.
More on the systems themselves at /voice-agents.
Study 02: A booking platform that ends the commission
Sector: hospitality and short term rental. Read it: /work/direct-booking-platform
Every reservation arrived through a listing marketplace that took a percentage of it, owned the guest relationship, and set the rules. We built the operator a full platform of their own: a booking engine and checkout, an operator portal, a guest portal, bilingual in English and Spanish, and live local data that refreshes itself.
What shipped:
- Booking engine and checkout
- Operator portal and guest portal
- Bilingual, English and Spanish
- Self refreshing local data
No client figures are published for this study. The operator has not consented to release booking data, and we do not publish numbers we cannot attribute. Their own description of the change: most bookings now arrive direct rather than through the listing sites, so the fee on those reservations stays with the operator.
Also in production
Running, but without a full study behind it yet, either because the engagement is young or because the client's figures are theirs to release.
- Search traffic that stopped being guesswork (online platform). Real search demand mapped and sized, a page and content architecture built around that map, on page data that updates itself, and answer engine work so ChatGPT, Claude and Perplexity can find the site and quote it correctly. Outcome: every page now aims at demand the team can name, watched in a dashboard that syncs itself daily. See /ai-search-optimization.
- A CRM built around how a trades business actually runs (trades and home services). Tracks every lead from first call to paid invoice: jobs, quotes, scheduling, and follow-ups in one place. Outcome: fewer dropped leads, faster quotes, a clear view of the pipeline. See /crm-automation.
- Books that keep themselves (operations and finance). An automated accounting pipeline that pulls transactions, categorizes and reconciles them, and produces the reports the business actually reads, with a human checkpoint where it counts. Outcome: reconciliation that used to take days, handled quietly in the background.
- The busywork, automated (across the business). Automations across sales, marketing, and operations that move data, trigger the right follow-up, and keep systems in sync. Outcome: hours back every week, spent on the work that needs a person.
Detailed numbers beyond what is published here are shared under NDA on a discovery call. Start with an AI audit or book a call.