Optiify
Predictive Maintenance
I helped build Optiify.ai, a building management platform that helps property managers and building owners identify, understand, and predict HVAC problems through predictive maintenance and an AI assistant called Optiimax. Our team designed the brand system, AI experience, and predictive maintenance platform. I led the brand direction, and helped turn early AI-generated concepts into a structured V1 dashboard and reusable component system that made complex HVAC fault data easier to understand and act on.
IMPACT
I built a cohesive brand and product direction for Optiify, including its logo, visual system, reusable interface components, and AI experience. The system helped unify a technically complex platform serving four distinct audiences with different levels of HVAC expertise. The foundational design was later adapted into a predictive maintenance concept for rail systems, extending beyond building management.
ROLE
Founding Product Designer
TIMELINE
3 months
Version 1 MVP
TEAM
1 Product Designer
1 Founder
4 Developers
SHIPPED
In Progress
THE PROBLEM
Buildings have many technical operations that leave behind complex logs full of data. None of it makes sense to the people running them.
A modern high-rise sends 60 data points every 15 minutes — temperatures, pressures, fault codes, air quality. That data goes to a dashboard. The dashboard goes unread because nobody knows what it means.
HVAC systems are one of the biggest financial drains in commercial buildings. When something breaks, property managers scramble to understand what happened, contact service techs without enough context, and react after the damage is done. The system exists to serve the building. The building ends up serving the system.
Optiify's vision: flip that. Give building operators a heads-up before failure happens — and give them the language to act on it. Starting with HVAC, but built to scale to every system in a building.
"You don’t want to let your building’s internal systems get to the 'what's that sound?' stage. As a building owner or as property management you want to be on top of your maintenance, but also not have to guess what’s going to break next and when."
— Ghalib, CEO of Optiify
I joined as the founding product designer to define the predictive maintenance and Optiimax, while establishing a consistent visual direction and design system with brand, color, typography, and interface patterns. I also helped flesh out three features that were in a temporary concepted prototype: Fault Detect. My role was to bring those pieces into a cohesive system ahead of the May 5 beta launch in Melbourne.
MY ROLE
THE AUDIENCE
The project was seen through the lens of 4 different personas. Those using, managing, fixing, and inhabiting are in a system.
Domain experts designed the system. Building operators have to use it. That gap is the design problem.
Two sides of the audience:
"Managing my properties’ HVAC are a headache. There's so much information and management and it'd be easy to have something just do it all for me or tell me in plain words, what’s going on and what I need to do."
— Airbnb Property Owner
"It's my dream to work in a high-rise. But if I have to stay late, I only have so long before fresh air runs out, and it just gets more and more stuffy. I have to leave before the lack of breathable air gets to me, and I can’t have important meetings that come up last minute in the building when it’s after hours."
— High Rise Employee who works after HVAC operational hours
These discussions helped me reframe the concept of After Hours, to not just be a scheduling feature. The goal is to make everyone’s jobs easier, whether as a tenant or property manager. All they need to do is send a request or handoff deciphered simplified information, without needing to be an expert.
MVP focus: Property managers, landlords, and tenants. Service tech flows were scoped for a later phase — their needs required deeper domain research we didn't have time to do properly.
THE RESEARCH
Competitive analysis showed that predictive maintenance tools were powerful, but still designed around technical users
The building management software industry has tried to give more data and similar AI capabilities but still relied on users to interpret what the problem meant and what to do next. None of them talked to building owners like normal people.
I also looked at how AI-native products handle conversation and context because we want Optiimax to act as a co-pilot instead of a chatbot.
I also looked at hardware context — Dyson, Ecobee, Siemens, and Honeywell sensors — to understand the physical layer that feeds the software. If a sensor is sending 60 data points every 15 minutes, the UI has to decide what 3 of those points actually matter to a building manager.
I also looked at hardware context — Dyson, Ecobee, Siemens, and Honeywell sensors — to understand the physical layer that feeds the software. If a sensor is sending 60 data points every 15 minutes, the UI has to decide what 3 of those points actually matter to a building manager.
WORKING WITH AI
Working with AI did not feel like a shortcut
Figma Make was used to concept screens early. It did the minimum — structure existed, text was placed — but the navigational logic was unconventional, the information density was overwhelming, and nothing had the critical thinking of actual design decisions. AI gave us a skeleton. The work was making it human.
DESIGN DECISIONS
Some changes I made and why
OUTCOME
What’s set to ship
MY REFLECTION
What I'd do differently
with more time and clearer direction.
This was fast, ambiguous, and cross-continental. Here's what I learned — and what I'd push harder on next time.
More time in each audience's world. I asked a high-rise worker and an Airbnb owner. That's a start — not enough. With more runway, I'd do structured research sessions across all four user types, especially service techs. Their needs were deferred from the MVP but they're the ones who close the loop on every fault.
The generation gap deserved earlier attention. Discovering that 60% of users prefer dashboards over chat interfaces happened mid-project. If that's foundational to product decisions, it needs to surface in week one — not week six. I'd build that research question into the very first sprint.
I would've tested more navigation patterns for Optiimax. I proposed tabs and side nav over buttons. I know that was the right call — but I'd want to actually test it. Not assume.
Ambiguity is the job — but that doesn't mean accepting unclear scope. I got better at designing with flexibility as things changed. What I'd do differently: ask harder questions up front about what's locked versus what's still in discovery, so I'm not designing for a moving target without knowing it's moving.
AI is a starting point, not a deliverable. The team's Figma Make concepts gave us something to react to — which has real value. But using AI output as a reference rather than a baseline changes the quality of what you build. I learned that early.