Meet Alfred

Inspiration

We didn't have to look far for the inspiration - we were already living the problem. Having to constantly switch between Gmail, Google Calendar, Apple Health, Notion, and a Todoist list with a long list of tasks, and last-minute deadlines. The overwhelm wasn't from a lack of tools. It was from having too many that never talked to each other.

Early in the hackathon, a quote really stuck with us: "I want AI to do my laundry and dishes so that I can do art and writing — not for AI to do my art and writing so that I can do my laundry and dishes." It perfectly encapsulated exactly what felt backwards about the world we were building in. AI was generating essays and writing code, yet we were still manually juggling our own daily chores.

Alfred grew from that frustration. and anxiety. What if there was one place that genuinely understood your entire life - not just your tasks, but your sleep debt, your energy levels, your step count, your upcoming deadlines - and turned all of that into a plan you could actually realistically follow?

What it does

Alfred is your personal assistant/manager. Every morning, fill in a 30-second "Daily Check-In": you tap one focus area (Academics, Fitness, Sleep, or Mental Health), and Alfred analyses your overnight sleep, current step count, upcoming calendar events, and recent mood logs to propose 5–6 smart, prioritised tasks, not some 50-item list, with the things that actually matter today. If you don't like a task, you can ask Alfred to swap it out for a completely new proposal too.

Throughout the day, the home dashboard tracks your progress in real time: steps tracked live, tasks get checked off, and upcoming assignments/deadlines stay front and center. Every week, a quick reflection closes the loop, allowing reflection on the highs and the lows, which feeds directly into the next week's suggestions. The whole experience is built around one principle: glanceable in 3 seconds, actionable in 30.

How we built it

We started with an ideation phase - mapping out the key areas of life we actually juggle every day: fitness, nutrition, sleep, academics, mental health. From there, we did a focused round of market research into each domain, exploring common pain points and existing feature patterns to sharpen our thinking on what Alfred genuinely needed to tackle versus what was nice-to-have noise.

After putting together project briefs and defining our core feature set, we sought out mentor feedback early - which proved invaluable in keeping us on track and gave us a way forward. With a clearer direction, we moved into wireframing, designing the key screens and user flows using Google Stitch and Figma to map out the experience before writing a single line of code.

Once the prototype was in good shape, we brought it to life using Google AI Studio, with the Gemini API serving as Alfred's brain - powering the personalised, context-aware task suggestions at the heart of the experience. From there, it was a cycle of building, debugging, and sketching out the next layer of features we're excited to bring to life.

Challenges we ran into

Like any project built under time pressure, we had our share of obstacles to work through. The first was conceptual: early on, we envisioned Alfred as a collection of specialized personas - a fitness coach, a sleep guide, an academic planner - each tailored to a specific life domain. It wasn't until a brainstorming session with our mentor that we caught the irony. We were recreating the exact fragmentation problem we set out to solve, just dressed up differently. That conversation forced us to step back, strip the idea down, and refocus on what Alfred was really about - one unified experience, not five smart ones.

The second challenge was on the implementation side. Alfred's core value depends on multiple data sources - calendars, health apps, step counters - all feeding into one coherent picture. Getting those integrations production-ready within the hackathon window wasn't fully achievable, so we're currently working with placeholder data while the framework sits ready underneath. It's less a blocker and more the natural next engineering step — the architecture is designed for it, we just need the time to connect the pipes.

Accomplishments that we're proud of

I think the biggest accomplishment was getting an actual proof of concept up and running. Watching Alfred go from scribbles on a whiteboard to a working proof of concept was a genuine "it's alive" moment - the kind that reminds you why you start these projects in the first place.

Beyond getting it running, we're proud of what it represents. The loop we've built - morning check-in, smart suggestions, and real-time progress - is something anyone can slot into their day, regardless of how packed their schedule is. It doesn't demand a drastic behavior overhaul. It just quietly makes your day easier.

We're also proud of how simple Alfred feels given everything happening under the hood. The 30-second morning ritual isn't a gimmick - it was a deliberate design decision we defended every time scope creep tempted us to add just one more step. Keeping that constraint was harder than it sounds, and we're glad we held the line.

What we learned

Building for yourself is a hack. When you're the target user, you stop second-guessing whether the problem is even real and just start solving it. That founder-user fit kept us grounded every time we debated a feature: we could just ask ourselves whether we'd actually use it.

We learned that data unification is genuinely hard - not technically, but philosophically. Deciding which aspects of someone's life deserve more weight requires a surprising amount of objective self-analysis. In trying to build a product that understands its users, we ended up learning a lot more about ourselves in the process.

We also learned that the right process makes everything clearer. Market research and wireframing weren't just box-ticking exercises - they actively shaped what Alfred became, helping us cut through the noise and zero in on what the product should actually prioritizes.

And maybe the most human thing we took away: feeling overwhelmed is often just a framing problem. Alfred doesn't eliminate the mountain of work - it just reminds you to take the next step, and then the one after that. Building it reinforced that lesson as much as using it does.

What's next for Meet Alfred

The immediate priority is real data. We want to start pulling in live context from users' calendars, emails, and health apps - closing the gap between what Alfred could know and what it currently works with.

From there, we want to deepen the intelligence layer: moving towards custom ML models that learn your patterns over time, so Alfred's suggestions get sharper the longer you use it. The goal is for it to genuinely know your rhythms, not just follow a rulebook.

Longer term, we're exploring university API integrations (like Canvas and Ed) to pull assignment feeds directly, Apple Watch support for richer health context, and a social layer for study groups to keep each other accountable. The ambition is simple: make Alfred the first thing you open every morning - and the reason your day actually goes to plan.