Repository navigation
MS3
Written by: Sofia Alcalá
For MS3, instead of running a brand new divergence-convergence session from scratch, we built directly on the results already produced during MS2's brainstorming process, using them as the foundation to evolve our selected solution (University Gastronomic Offer) with smart features.
We went back to several post-its from our MS2 board that had strong potential to become smart features and asked "what would the intelligent version of this idea look like?" This produced the following smart features, each grounded in a specific idea already generated and categorized during MS2:
- Restaurant crowd level: a real-time indicator of how crowded a place is, built from user-submitted reports, plus estimated wait time and push notifications when a saved favorite's crowd level drops.
- Chatbot to help choosing restaurant: a conversational assistant that asks the user a couple of quick questions (budget, cravings, time available) and returns a ranked suggestion instead of making the user browse manually.
- Custom dietary preferences saved in the user's profile: a recommendation engine that automatically filters and ranks results based on the user's saved profile (vegetarian, halal, gluten-free, etc.) without the user having to re-apply filters every time.
- Filter by available time between classes: a schedule-aware suggestion feature that, given how much time the user has before their next class, only recommends places they can realistically visit and return in time.
- Group voting to choose where to eat: a real-time group decision tool where multiple users vote simultaneously and the app surfaces a consensus recommendation, useful for groups of friends who can't agree on where to go.
- Live daily menu / today's specials: a dynamic feed of daily specials pulled directly from partner restaurants, updated automatically instead of relying on static menu photos.
Because these ideas had already gone through convergence in MS2, we used that existing prioritization as our starting point: the crowd-level indicator and the dietary-preference features had already surfaced as high-value ideas, while the chatbot, schedule-aware suggestions, and group voting had remained as lower-priority divergence ideas. We kept that same relative prioritization when translating them into smart features: the crowd-level indicator and the personalized dietary recommendation engine became our top two smart features to build first, since both directly respond to needs explicitly raised in our interviews (Karim's frustration with crowded restaurants, and Ana María's dependence on vegetarian-friendly options), while the chatbot, schedule-aware suggestions, and group voting tool remain as stretch goals for later in the semester.
Written by: Sofia Alcalá
The team's deciding criterion was the reach of the problem across the university community. Unlike Parking (which only affects those who own a vehicle), Campus App (which mainly helps newcomers), or Second Hand Clothing (an occasional, not daily, need), the question of where to eat is something every single member of the community faces, regardless of what semester they're in, everyone needs to decide where and what to eat near campus every day. This gives the problem a sustained, universal impact, meaning any solution we build generates real, consistent value for the largest possible number of users, rather than just one specific segment of the student population.
After deciding, we relied on the divergence-convergence pattern taught in class, which we had already applied once during MS2 and revisited here as the foundation for our final decision.
Divergence: During the Ideate phase of MS2, the team generated ideas without filtering or judging them, following the "yes, and..." rule instead of "yes, but...", which encourages building on a teammate's idea instead of shutting it down. We wrote over 40 ideas from simple features (search restaurants by name, photos of the place, save favorite places) to more disruptive ones (a chatbot to help choose a restaurant, a "surprise me" button, a "lunch roulette").
Convergence: Once the divergence phase was closed, each team member cast three votes across all the generated ideas using post-it voting. The most-voted ideas were then sorted into the four convergence categories taught in class:
-
The Rational Choice (most feasible with current resources): centralized content, search by name or by food type, nearby restaurants, opening/closing hours, photos, saved favorites, and location assistance.
-
The Most Likely to Delight (what would most engage users): friends' recommendations, helpful reviews, a "popular this week" section, and daily promotions.
-
The Darling (the team's personal favorite): "lunch roulette," a decision quiz, and a cravings-based survey.
-
The Long Shot (high risk, high reward): real-time restaurant crowd level, delivery partnerships (e.g., Rappi), and pickup options.
After all that, we confirmed that this would be the idea to work on, reasuring that this idea is the option most aligned with the broad-impact criterion.
After conducting two interviews, we found the first prospective user: a member of the university community who invest time and effort in doing research before choosing where to eat near campus. They check reviews on Instagram or TikTok and is influenced by a place’s popularity or recommendations from friends, but they discover that this preliminary research doesn’t guarantee a good experience: as happened to Karim, who arrived with high expectations at a spot very popular among students and ended up disappointed because “the burger wasn’t that good,” or as Ana María describes it more subtly, acknowledging that even when she’s seen very good reviews, she sometimes ends up disappointed and admits that “it’s not so certain that the effort was worth it.” This user experiences a constant tension between the effort they put into gathering information and the real uncertainty of whether that information will translate into a good decision, which gradually erodes their trust in the sources they use to choose where to eat.
We identified a second prospective user: a student with very limited time between clases who prioritizes speed, proximity, and predictable when choosing where to eat. Karim mentioned that she sometimes has only around thirty minutes between classes, while Ana María also tends to repeat familiar options when time is limited. This user is mainly affected by the uncertainty of waiting times and distance, which often leads them to choose nearby, known restaurants instead of exploring new alternatives.
We identified a third prospective user, a student with specific dietary preferences who looks for balanced, affordable, and suitable food options near campus. Ana María, for example, follows a vegetarian diet and often returns to familiar restaurants because she already knows they offer meals that fit both her nutritional expectations and budget. Although there is a wide gastronomic offer around the university, this user faces fewer realistic options once dietary needs, price, variety, and available time are considered together, which often leads them to rely on the same trusted places instead of exploring new alternatives.
We identified a fourth prospective user: a student with a limited food budget who wants to find affordable options near campus without sacrificing quality or variety. This user usually compares prices, checks menus, looks for promotions or lunch specials, and often returns to familiar places where they already know how much they will spend. However, prices are not always clearly available online and promotions can change frequently, making it difficult to know which option offers the best value. As a result, this user may feel frustrated or limited by their budget and tends to prioritize restaurants with predictable prices and a good price–quality balance instead of exploring more expensive or unfamiliar alternatives.
We identified a fifth prospective user: a student who usually eats with a large group of friends and struggles to find a place where everyone agrees and can sit together. This user experiences the constant tension of balancing different cravings, individual budgets, and the physical space available at nearby restaurants. The challenge of group decision-making often wastes a significant portion of their short break time, which frequently leads them to compromise on their preferred food or settle for the same familiar spot just to avoid further arguing and ensure the whole group can actually eat together.
We identified a sixth prospective user: a student on a strict daily budget who wants to find affordable, filling meals near campus without sacrificing quality. This user is highly sensitive to the elevated prices of many popular spots and the hidden costs of eating out, constantly searching for daily specials, combos, or student discounts to make their weekly allowance last. They are mainly affected by the financial stress of dining around campus, which often leads them to walk further away to find cheaper alternatives, skip buying drinks, or feel excluded when their friends choose more expensive restaurants.
Writen by: Sofia Alcalá and Juan Felipe Ochoa
The solution addresses University Gastronomic Offer: despite the wide variety of places to eat, drink coffee, or grab a snack around Universidad de los Andes, there is no single centralized place where students and staff can discover, compare, and decide where to go. As confirmed through our interviews, people currently rely on a fragmented mix of Instagram and TikTok accounts, word of mouth, a process that is time-consuming, inconsistent, and doesn't account for personal constraints like dietary restrictions, budget, or the short time windows between classes.
Our proposed solution is a mobile app that centralizes the university's gastronomic offer into a single, easy-to-navigate experience, structured around six core screens validated in our low-fidelity prototype:
• Home Discover: the homepage acts as the main discovery space of the app, presenting personalized restaurant suggestions, nearby options, popular places, and relevant categories so users can quickly explore the gastronomic offer without needing to search across multiple external sources.
• Campus Map: an interactive map showing all nearby eating spots, whether they're currently open or closed, and a list of nearby places with estimated wait times, directly addressing the lack of centralized, location-aware information.
• Search: a dedicated search option allows users to look for restaurants, specific types of food, or categories and apply relevant filters such as price range, dietary preferences, and distance, making it easier to find an option that matches their immediate needs.
• Place Detail: a dedicated page per restaurant showing photos, hours of operation, location with directions, and general information, giving users the confidence to make a decision without needing to leave the app to verify details elsewhere.
• Saved & Food Preferences: users can save favorite spots, filter by dietary preferences (vegan, gluten-free, halal, vegetarian), and complete a short onboarding survey about the cuisines they enjoy, so recommendations are personalized from the very first use.
• Preferences survey: a short survey presented during onboarding gathers information about the user's preferred cuisines, dietary restrictions, budget preferences, and other relevant food interests. This information is then used to personalize the content shown across the app and make restaurant recommendations more relevant to each user.
Beyond this core solution, we plan on incorporating a set of smart features derived directly from patterns identified in our interviews and refined through our team's divergence-convergence process from MS2:
• A real-time crowd-level indicator, letting users see how busy a restaurant is and its estimated wait time before heading there, responding directly to the frustration described by one interviewee about all students going out to eat at the same time.
• A personalized dietary recommendation engine, automatically filtering and ranking results based on the user's saved dietary profile, addressing the recurring difficulty faced by users with specific dietary restrictions who must repeatedly search for the same limited set of compliant options.
• Planned stretch features for later in the semester include a conversational assistant to suggest a place based on quick prompts (budget, cravings, time available), a schedule-aware suggestion feature that only recommends places reachable within the user's free time between classes, and a group voting tool for friends deciding together where to eat.
Together, these features turn a currently fragmented, word-of-mouth-driven decision process into a single, personalized, and time-aware experience, reducing the uncertainty and repeated disappointment our interviewees described, while helping smaller or lesser-known establishments gain visibility within the university community.