We ask every client for a candid review six months after deployment. These are unedited excerpts. If you want to speak with any of these companies directly, we are happy to arrange a reference call.
We had tried two other vendors before Assured Smart. The difference was that their team spent the first two weeks just listening and mapping our warehouse processes. The demand-forecasting model they built cut our overstock by 22% in the first quarter. Our purchasing manager now checks the dashboard before every order.
The document-processing pipeline handles about 1,200 invoices a week for us now. Before, that was a full-time job for three people. Two of them moved into client-facing roles they actually enjoy, and the third retired on schedule. Accuracy is higher than manual entry ever was.
I gave four stars instead of five because the initial timeline slipped by about ten days. That said, the team was transparent about the delay, and the final product works exactly as promised. Our patient intake predictions are accurate enough that we reduced wait times by 18 minutes on average.
We are a 40-person marketing agency and our data literacy was, frankly, low. Assured Smart built a reporting dashboard that pulls from Google Analytics, our ad platforms, and our CRM. Now every account manager can answer client questions with real numbers during the call instead of saying they will check later.
We manufacture custom cabinetry. The quoting process used to take our estimators about 45 minutes per job because they had to look up material prices, labour estimates, and finishing options in separate spreadsheets. The tool Assured Smart built pulls everything into one screen and generates the quote in under five minutes. We close deals faster because we can reply the same day.
I was sceptical about AI. It sounded like a buzzword. But the Assured Smart team showed me a working prototype within three weeks, using our own sales data, and the predictions were close enough to be useful immediately. We have since expanded the project to cover inventory planning as well.
Groupe Véloce moves freight across Quebec and Ontario with a fleet of 85 trucks. Their accounting department received invoices in PDF, email body text, and even photographed paper documents. Three full-time clerks spent most of their day copying numbers from those sources into their accounting system.
We built an OCR pipeline that reads each incoming document, identifies the vendor, extracts line items and totals, and writes a draft entry into their Sage 50 instance. A human reviewer approves or corrects each entry before it posts, which takes about 15 seconds per invoice on average.
The project took nine weeks from kickoff to production. During the first month live, the error rate was 2.1%, compared to the 4.7% error rate the manual process had produced in the previous year according to their internal audit.
Clinique Santé Plus operates three walk-in clinics in the greater Montreal area. Their biggest complaint from patients was unpredictable wait times, especially on Monday mornings and Friday afternoons. Staffing decisions were based on gut feeling and last year's calendar.
We trained a time-series model on 28 months of visit records, cross-referenced with weather data, public holidays, and local event schedules. The model predicts next-week patient volume per clinic per half-day slot with about 88% accuracy.
Clinic managers now receive a staffing recommendation every Thursday for the following week. They can adjust it, but in practice they follow the model about 80% of the time. Average patient wait dropped from 47 minutes to 29 minutes within four months of deployment. The clinic hired one additional nurse for peak slots and reduced overtime costs by roughly $4,200 per month across all three locations.
Tell us about your challenge and we will let you know, honestly, whether AI software is the right tool for it.
Talk to us