FairTime

Inspiration

The topic of Max Ease joined with Wise's product pillars of speed, convenience, and transparency led us to rethink group scheduling from the ground up. We realised that scheduling shouldn't require hidden personal sacrifices and so we built FairTime to move away from slow, awkward negotiations and hidden compromises toward an automated, transparent, and mathematically fair experience.

What it does

FairTime is like Grammarly but for fair meeting scheduling. Rather than merely identifying blank calendar slots, it calculates the personal inconvenience of meeting times for each attendee, balances sacrifices across the team, and suggests the fairest option with complete mathematical transparency. It acts as an ambient extension that overlays your active workflow from your current digital calendar (such as Google calendar or Outlook). Rather than forcing groups to switch to a third party polling site or dashboard, FairTime sits along your existing tools as an extension that can be used as required when scheduling meetings.

How we built it

Under the hood, FairTime operates as a fully deterministic system rather than relying on blackbox AI. Our engine takes in authorized calendar free vs busy data, automatically filtering out protected windows and existing commitments such as rest times and assignment deadlines. It then evaluates feasible slots in 5 minute increments, scoring penalty costs based on distance from each attendee’s preferred working hours. To ensure no single person bears the brunt of the compromise, FairTime runs a minimax optimization algorithm to minimize the worst case individual burden across the group. Finally, an ACID-compliant time series ledger on Tiger Data PostgreSQL looks at each user's historical inconvenience, ensuring past sacrifices are remembered and team members who took late calls previously are automatically protected in future sessions. The frontend delivers this seamlessly via an ambient Manifest V3 browser extension embedded directly inside Google Calendar and Outlook, paired with a Next.js web application for account and group settings.

Challenges we ran into

With tour entire team coming from non-CS backgrounds, coordinating multi-branch Git merges, configuring relational schemas in Tiger Data (PostgreSQL), and structuring browser extension messaging under the deadline was intense and stressful. To bridge this gap, we used generative AI tools like GitHub Copilot and ChatGPT as interactive pair-programmers to help us navigate unfamiliar syntax, debug complex problems, and scaffold our Tiger Data PostgreSQL schemas. Every line of generated code was manually reviewed, integrated, and validated by our team.
Because we worked in parallel across distinct components - the browser extension UI, calendar ingestion, minimax scoring engine, and Tiger Data PostgreSQL database - linking everyone’s individual modules into a unified, seamless pipeline was a significant engineering hurdle. Enforcing strict API data contracts and standardizing UTC ISO timestamps early on proved essential to prevent integration breaking bugs during the final assembly. Ultimately, we found that the most effective solution was the most simple - open, constant communication and regularly syncing on what each person was tackling and how we were approaching it proved to be the most useful tool.

Accomplishments that we're proud of

We translated Wise's core product pillars (speed, convenience, and transparency) into a direct solution for an everyday problem that we ourselves face and then successfully engineered a browser extension that allows users to request, inspect, and confirm meetings without manual availability entry. Which we did within the tight time limit and working with completely new technical territory.

What we learned

-When our code broke while linking the modules together, the solution didn't come from writing more code. It came from active team communication stopping to walk through what each person was building and agreeing on strict API contracts before merging. -As first time builders, we learned how to use generative AI responsibly. We used tools like ChatGPT as an interactive tutor to explain unfamiliar syntax and query structures, while keeping ownership of our architecture. -Solving an issue we live through weekly gave us clarity under pressure. Beyond rapidly expanding our technical abilities, collaborating on FairTime accelerated our personal growth as it sharpened our communication, collaborative problem solving, and presentation skills.

What's next for FairTime

Looking ahead, we plan on scaling FairTime from focusing on students to distributed enterprises. If FairTime reclaims just 35 minutes per person of wasted scheduling coordination time per week, a 10 person team can reclaim roughly 300 hours of wasted time each year. This translates to approximately £15,000 in saved time annually. Not only does FairTime prevent worker burnout (by scheduling fair meeting times), it also boosts team satisfaction which further boosts productivity.