Michael Marelli, Founder of FilterSync by ResProof, has published a review on Clutch.
Michael rated AnyforSoft 5.0 across all categories: quality, schedule, cost, and willingness to refer. On NPS, Net Promoter Score, which is the likelihood of recommending a partner to others, Michael gave us the maximum score.
The review marks the end of the first chapter of a collaboration that continues to develop.
A platform built to prevent $6K–$12K HVAC repair bills
FilterSync started as a problem with no existing solution to reference.
Property managers had no reliable way to confirm that tenants had changed their HVAC filters. Missed changes accumulate over time. The cost shows up later, in the repair bill.
AnyforSoft designed every part of the platform from scratch.
The team built:
- A mobile-first tenant submission flow triggered by a single SMS link
- An AI verification engine that analyzes each photo for authenticity, timing, and filter condition
- A human-in-the-loop escalation layer that handles edge cases without slowing the system
- Role-based dashboards that give property managers portfolio-wide compliance visibility in real time
- A modular architecture designed to support additional compliance categories beyond filters
The platform launched in 2025.
Today, FilterSync has early market traction and a provisional patent – the first formal step toward protecting an original invention.
“When stuff came up (and stuff always comes up), they didn’t just dump problems in my lap. They laid out options, explained the trade-offs, and we moved forward together.”
Michael Marelli, Founder at FilterSync
First iteration: 95%. The next milestone is higher
The collaboration did not end at launch. The collaboration continues. AnyforSoft keeps developing FilterSync as Michael’s vision for the platform grows.
One active focus is AI accuracy.
“We reached 95% verification accuracy in the first iteration. That is a strong result for a computer vision system operating across inconsistent lighting and device conditions. Now that we have more data, we are working to improve that metric further. AI performance is not fixed at launch. It develops as the system accumulates real-world data, and that is worth keeping in mind when planning an AI project.”
Oleg Bogut, CTO at AnyforSoft
The case study tells the full story: the challenge that started it, the architecture that holds it together, and the verification loop that makes it reliable.