top of page

A day in a life

What It's Really Like to Work at Northwood

Modern office workspace with iMacs, desk lamp, and plants

Getting to Know SunExpress IT Airline Business Solutions: One Team, 21 Products and the Story of CrewBuddy

Yusuf Burakhan Sert
17 minutes ago
18 min read

Where the pulse of operations beats 

When I joined SunExpress's IT Airline Business Solutions department four years ago, I didn't fully realize I was stepping into such an exciting place, one where the pulse of operations beats. It turned out there was so much more to it. Today I'm the PO of the Crew Solutions team. We build products focused on cabin and cockpit crews, and we support more than 20 applications owned by our team alone, all at the same time. Each has its own active needs, current requests and areas to develop. That requires knowing the monthly and daily work of cabin and cockpit crews and office staff really well. A delayed flight, a rotation change or the last day of a training: each comes with its own excitement, pace and priority. To truly understand this world, our first move was to map every process of our cabin and cockpit crews end to end as a journey map, from planning the work month to flying, from the layover to returning to base. At every step we marked which programs they used and which tasks were still done manually. That showed us exactly which features could be added to which product and where the manual touchpoints were. While building the journey, we also joined test flights at different times and observed the process on site, exactly as it happens.


Crew journey map: every stage from planning the work month to finishing the trip
Crew journey map: every stage from planning the work month to finishing the trip
Journey map close-up: each step and the system used at that step
Journey map close-up: each step and the system used at that step

This work also changed our relationship with the business units. Before, they would bring an idea and we would build exactly that. With the journey map, empathy became easier; we started interpreting requests, adding an IT perspective and generating product ideas with greater added value and scalability. Our goal was to reach a level of knowledge where, even though we are POs and not the business unit, we could do their job the same way. Of course we can't fly the aircraft, that's not what I mean 🙂 But we wanted to understand operational processes and their meaning just as deeply, master the aviation domain end to end and speak the same language as the business units. Today, Crew Solutions manages 21 products with five developers and actively serves six different departments. And we do it with high user satisfaction. As POs, we were wearing many hats long before AI entered the conversation: analyst, designer, tester, support, product owner and product manager. Yes, none of these roles exist separately on the team. We take them all on as POs and drive the processes this way. To wear all these hats at once, and as efficiently as possible, we had to discover new methods. And we did; I'll cover them in upcoming articles in this series. The newest of these methods, and to me the most exciting, is AI. Our AI agents now wear a large part of the analyst, designer, developer and tester hats: from the moment a request arrives, they can run the entire process, from analysis and design to development, testing and deployment to the test environment. They pick up the ticket in Jira themselves and debate with a review agent to get the code ready; in sprint planning, we've started assigning stories to AI in Jira, just as we would to a developer. We still check everything before it goes live, but honestly, finding issues is getting harder and harder. I'll go into the details later in this article. To get here, we went through many cultural and radical changes and tried dozens of governance methods. We measured all of them. What worked became a permanent part of our process; what fell short of expectations we didn't hesitate to drop. We embraced agility not as a methodology but as a habit. I want to sincerely thank both my managers, who made room for this transformation, and my team, who stood shoulder to shoulder at every step.


What's coming in this series? 

This is the first article in a series. Throughout the series, I'll share which products we build for cabin and cockpit crews, how we manage them, which operational topics we deal with, and where we stand compared to similar applications in the industry. My aim is to share real knowhow with anyone interested in the topic and with my peers across the industry. 


One development, four or five times the impact 

Every product we build consists of hundreds of pages. Fellow POs will know exactly what that means in terms of workload and demand. That's why we never treat a development as the answer to a single request. We build dynamically, analyze edge cases upfront, solve problems at the root rather than the symptom, and cover likely future requests today. We make one development but aim to multiply its impact by four or five. Let me give a concrete example: we used to receive at least 2–3 support cases a day from the Survey module. We analyzed the module from scratch, identified every edge case and redesigned the system from the ground up. We convinced the business unit of this approach and took it live. Today we receive almost no cases from this module; maybe one a month, if that. This single development significantly reduced the support load in that area. We have to work this way. How else could five developers build more than 20 applications in parallel, with the same level of quality? 


The apple of our eye: CrewBuddy 

Throughout the series we'll talk about every application to some extent. But there's one that holds a special place for us: CrewBuddy. It's the first application we built end to end with Scrum, and it's entirely in-house. CrewBuddy is the most innovative and modern product in our portfolio. It's ahead of many of its counterparts on the market and offers many features they don't have. I'm not saying this lightly. We regularly analyze competing applications and closely follow the features they offer. We try to learn as early as possible not only what exists today but which features are coming in the short and medium term. If a high-value feature isn't in our backlog, we run a needs analysis and add it to our product backlog. We started building it four years ago so crew members could manage all their work in one place, and took it live two years ago. Today it's the most actively used application in our portfolio and has become an indispensable companion for our flight crews. Our backlog is also in very healthy shape. Let's look at a single day. 4:30 a.m. In Antalya, a cabin crew member turns off the alarm and picks up the phone. Today's flight, who they'll fly with, the aircraft rotation and the tasks ahead are all on one screen. They can check a colleague's roster to plan accordingly, and follow their training information and details from the same place. They see the passenger information they need for cabin preparation down to the finest detail, before even stepping onto the aircraft. In the same minutes, a captain checks the aircraft status reports and information about the destination. If there's a document to read and acknowledge before the flight, a notification lands directly on their phone. When they want to reach their crew or operations departments, they don't switch to another app; they can message them directly through CrewBuddy, including the operations control centers. Jump seat requests that used to run on email threads are now approved in two taps. Both open the same app, but each sees an experience personalized to their role. On top of that, everyone can choose the modules they want to see and customize their home screen. And most of this experience is shaped not by code but by a fully dynamic admin panel. The operations staff in the office: 


  • Send notifications to exactly the right people, filtered by user, base, flight, rank and many other parameters. 

  • Build surveys themselves, from the questions to the weighting calculations and transcript publishing. 

  • Create forms themselves from scratch. They choose when, to whom and where in the app a form appears and publish it according to the rules they set. When a new process needs to be digitalized, there's no need to wait for a development specific to that form; the process moves forward without waiting in IT's backlog. And when a form request comes in, we build it to cover the different form types that may come later. This way, the impact of a single development multiplies. 

  • Monitor operational events through tracking modules and intervene. A crew member removed from a flight, a change, an aircraft diversion; every scenario is managed from the panel. 


All without writing a single line of code. 

At many airlines, this scene still looks very different. The information flight crews need is scattered across roster systems, corporate portals, PDFs, email chains and WhatsApp groups. Every new need adds a new app, a new password, a new form. The result: the industry's most valuable resource, its crew, spends its time searching for information. 

You might think an application at this scale is backed by a huge software organization. It isn't. CrewBuddy and the other 20 products in our portfolio are built by the same small team. 

So how does a small team move this many products forward without them crowding each other out? First, let's get to know CrewBuddy a little better, then move on to our product decisions, our prioritization model and our team culture. I believe there are transferable lessons here not only for people working on crew experience in aviation, but for any product and technology leader managing a broad portfolio with limited resources.


CrewBuddy: The operations center in a crew member's pocket 

CrewBuddy is a super app we designed for more than 2,000 cabin crew and more than 1,000 cockpit crew. But it's not only for crew who fly: Cabin and Cockpit Operations, Training Management, IOCC and Technical teams use the same platform. What's inside the app?


  • Roster: All duty types, roster sharing and a personalized view.

  • Flight details: Aircraft rotation and crew information; detailed passenger information for cabin crew, aircraft status reports and destination information for cockpit crew.

  • All flights: Every SunExpress flight, in full detail, in one place. 

  • Duty Chat and communication: Direct communication between crew members on the same duty

  • IOCC communication channel: Direct communication with the operations control center from within the app. 

  • Quick links: One-tap access to frequently used tools; each user can customize their own list. 

  • Docs: All the documents the crew needs, in one place. 

  • Cabin evaluations: Peer evaluation, scorecards and cabin surveys. 

  • Commission entry, reports and statistics. 

  • Jump seat approvals: One tap from the captain's interface


The flight details screen is a good example of this approach. Everything about a flight comes together in three tabs:


  • Flight Info: STD/ETD, STA/ETA, OFP number, aircraft registration, crew composition, expected DOW/DOI and the aircraft's current A/C Status report. 

  • Crew Info: The role and contact details of every crew member on the flight, one-tap call or message, and passport and ID validity status. 

  • Pax Info: Total passengers, INF, WHC, UM, XLEG, SPH, VIP, care list, pre-orders, jump seat passengers and sales target.


Destination Line brings the destination airport into the crew's pocket. Alongside a route map and quick facts such as weather, electricity, time zone and language, it covers transportation, station contacts, operational, hotel and security information, a pronunciation guide, announcements, product and service requirements, and entry and customs forms. None of this content is hard-coded. Headings, entries, when they appear, whether they're active or inactive, and their notifications are all configured by office staff from the admin panel. So when a new base is added, they can manage the entire process themselves without opening a ticket for the software team. The most used features, in order, are My Roster, Flight Details, My Actions, Cabin Survey, Cabin Commission and A/C Rotation. In other words, people use the app not as an "information screen" but as the place where they get their daily work done. In my view, CrewBuddy's real strength is on the side you don't see: a fully dynamic, no-code admin panel. Cabin and Cockpit Management manage content, announcements and forms themselves. Not every new survey or announcement lands on the software team. For a small team, this means capacity stays focused on product development.


Listening first: what 1,132 people told us

Drawing a roadmap at a desk is easy. We chose to ask our users instead. Our NPS survey started with 342 people in the first round and grew to more than 1,132 in later rounds, with responses from cabin (CCM, SCCM) and cockpit (CP, FO) crews. 

Here too, we cared about scalability. Instead of a one-off survey, we built a structure that both we and the business units can use whenever we want. We can create dynamic, rule-based NPS surveys and send them to exactly the right user group. We can even take the pulse of users before taking a feature live. This way, we listen to our users continuously, not in a once a-year survey. 

The results told us two things:


  • Ease of use is our strongest point (about 4.2 out of 5). Our design decisions were in the right place. 

  • Speed and performance were our weakest point (about 3.8). This was a clear assignment for us. We made our plans and analyses, focused the first quarter of the year on it, built the improvements and took them live. Of course, we then sent the survey again and saw this score rise radically.


About 35% of open-ended answers pointed to the same request: statistics and duty hours. We also saw that the two segments with the lowest satisfaction (SCCM and FO) had completely different problems. Had we built for one "average user," we would have missed both. For us, a survey isn't a report card; it's an input for prioritization.


An example: Jump Seat, from email to two taps 

Until June 2026, jump seat requests ran on email. Request, approval, tracking, reminders: everything got lost in inboxes. We didn't solve this in a single application; we connected two of our products. The request is created in SunPass, our employee ticketing system, approved in the captain's interface in CrewBuddy, and a push notification goes out at every step. We delivered this together with the Automation & Innovation team and Cockpit Management. The lesson for me: if your portfolio is broad, the greatest value sometimes lies not in a new product, but in the gaps between existing ones. Managing 21 projects can look like a burden; but it also means you're the only team that can build these connections.

Product decision: Why one door instead of ten apps?

In aviation, crew software usually grows around purchased core systems: roster, training, documents, reporting. Each has its own interface, and users jump between them. We chose a different path. 


  • One front door, many back-end systems. We didn't try to replace the core systems. We gathered operational data (flights, crew, ground handling) into a data layer and built CrewBuddy on top of it. Users don't need to know which system runs in the background. 

  • Task-focused, not role-focused design. A cabin crew member, a captain and an IOCC employee use the same app; each sees an experience shaped by their task. 

  • Business unit ownership. Thanks to the no-code panel, content is owned not by IT but by operations itself. This brings both speed and ownership. This choice can be applied the same way at another airline with a different set of systems. The real decision isn't technological: it's deciding whether you show users your systems or their work.


A closer look at the modules

I wanted to make what I've described a little more concrete, module by module. 

Home screen: An experience that changes with the time of day The home screen greets users with a theme that changes with the time of day. The next duty, one-tap access to IOCC and Destination Line, announcements and personal shortcuts are all on the first screen. 

Dark mode and Duty Chat: With night flights and early briefings in mind, the entire app switches to dark mode with a single tap. In My Actions, every duty card gives direct access to filling in surveys, entering commission and Duty Chat. Duty Chat is an end-to-end encrypted group chat that automatically brings together the whole crew on the same duty. 


Example from the CrewBuddy app homepage
Example from the CrewBuddy app homepage

Everyone's own home screen: Users decide which shortcuts appear on the home screen and in what order; they add, remove and reorder them. On top of that, they design how the modules look entirely by themselves, without any code. 


Roster and duties: The roster shows all duty types, color-coded; with My Buddies, users can view their colleagues' rosters. On the My Actions screen, past and upcoming flights, surveys, commission, Destination Line and Duty Chat are on the same card. 


My Actions: Every action in one place My Actions gathers every action the crew needs to take under the flight cards. Each action, such as filling in a survey, entering commission or Duty Chat, opens its own page and its own flow. Whenever a new action is needed, we can add it here too, so the crew completes everything they need to do from a single place. For example, tapping Fill Survey opens the list of crew members on that flight, and each evaluation is completed within the same flow.


Flight details: From the flight card, reaching IOCC or opening a jump seat request is one tap. Flight Info includes the OFP, registration, crew composition, DOW/DOI and the aircraft's A/C Status report; Pax Info includes every passenger breakdown needed for cabin preparation. 


Crew information: Each crew member's role, contact details and document validity are on one screen; a call or message is one tap away.


IOCC and operational tools: From the Contact to IOCC screen, crew can report situations such as a missing ramp agent, pushback or fuel truck with a single tap. This process used to run by phone; there was no visibility into how quickly a request was answered or whether the issue was resolved. Before building the feature, we defined the units' response targets as KPIs. By delivering system-level improvements alongside it, such as notifications, criticality levels and a tracking dashboard, we made the feature sustainable. A/P Timetable shows departures, delays, stand positions and the inbound aircraft by airport; Aircraft Rotation shows the aircraft's rotation and crews throughout the day. 


Cabin commission: Cabin crew track their sales and commissions by day and by flight, and enter their own entries step by step within the app. 


Evaluation and transcripts: Periodic transcripts, score trends and evaluation criteria are transparently in the crew's hands. 


Docs, Quick Links and other tools: Documents, links to frequently used systems and all crew tools in a single menu. In Quick Links, every user builds their own favorites. 


Notifications and profile: Targeted notifications and announcements, together with read/unread tracking. The profile offers dark mode, biometric login, in-app feedback and OTA updates that arrive without waiting for the app store.


Six departments, one team: Prioritization is a shared decision, not a negotiation

Alongside CrewBuddy, TMS, EFB, CabinSuite, CockpitSuite, SunPass, e-Grading and more are owned by the same team. With these products we actively serve six departments. Each has different development requests, urgent issues, and regulatory and audit-driven feature needs. Let's be honest: looking at demand and resources, we can't take on everything. But I believe we've found the secret to keeping stakeholders satisfied even when we can't. The secret is not making the decision alone.


1. Annual portfolio: Drawing the big picture at the start of the year 

At the beginning of each year, we collect annual portfolio requests from the departments and prioritize them. That way, every new request during the year is weighed not against a blank page, but against a plan everyone knows. I'll go into detail in upcoming articles about what we base prioritization on and how our scoring structure works. For now, let me say this: requests coming from outside the portfolio during the year are usually above 70%. In other words, we face a wide variety of requests far, far beyond what was planned at the start of the year. How we adapt planned resources to this variability will also be the subject of that article.


2. Monthly prioritization: Everyone at the table, everyone a decision-maker

Every month we meet with stakeholders from all six departments and challenge each department's priorities with everyone in the room. We're not the only ones deciding; the other five departments are decision-makers too. While one department argues for its request, the others explain why their own work can or can't wait. We convince each other. We leave the meeting with a clear list of focus topics until the next month. Even the owner of a request that didn't make the cut takes ownership of the outcome, because they saw how the decision was made. Transparency makes "no" acceptable.


3. Analysis the moment a request arrives: Measuring value in a shared language 

The moment a request arrives, we start analyzing it and evaluate it against shared criteria: • Financial gain: Does it bring direct revenue or savings to the company? 


  • Cost reduction: Does it eliminate or reduce an existing cost? 

  • Man-hour savings: How many hours of manual work does it eliminate in operations or the office? 

  • Disruptive innovation: Does it have the potential to fundamentally change how the work is done? 


We read this value together with the size of the work. A high-value but very expensive item can fall behind a medium-value, lightweight one; what we look at is the value produced per unit of effort. Regulatory and audit-driven work, on the other hand, goes to the front without scoring. Tying this to a written rule saves us from having the same debate every month.


4. Clear requests, fast development 

What speeds up development the most is neither a new tool nor overtime. It's a request that's clear and mature enough not to change radically during or after development. Small changes happen, of course; managing them is a natural part of the process. But pulling a request into a sprint before its analysis is complete is the most expensive mistake a small team can make. That's why every item that enters the team does so with a clear understanding of what is being asked and why. 


5. Embedding repetitive work into the product 

If every new announcement, form, or survey landed on the software team, meeting the demand of six departments would be impossible. The no-code admin panel exists for exactly this reason: routine work is in the stakeholders' own hands, while development capacity is reserved for what's genuinely new.


6. Building a platform, then building on it

Once the notification infrastructure, data layer, and shared components are set up correctly, a new module doesn't start from scratch. That's how we were able to take Jump Seat live so quickly. Every new product builds on the one before it. None of these is revolutionary on its own. But applied together, prioritization stops being a negotiation decided by whoever is loudest. It becomes a decision made by six departments together, openly and with clear reasoning.


Team: Flexibility, not military discipline 

The easiest mistake in a small, multi-project team is forcing everyone into a single rhythm. We tried the opposite. Listening to people, giving them room, and seeing who shines in which problem delivered more than a rigid process ever could. This flexibility doesn't mean a lack of discipline. According to our sprint data, in the first half of 2026 our story point completion rate was 97.6%, and our sprint goal achievement rate was 91%. But what we really look at isn't output; it's outcome: what the product changed in the user's life.


What about AI?

I want to be honest about this: this team model was working before AI. AI didn't give us a capability from scratch; it added leverage to a system that was already working. Today we are embedding AI agents across the development lifecycle, from analysis to testing, and we use so many different methods that it's impossible to fit them all into one article. For example, AI can now develop the support tickets we receive as well as some R&D developments. When I say "AI develops," I don't mean a developer writing code with AI. From the moment a request arrives, AI runs the process:


• It analyzes the request and defines the requirements.

• It designs the interface according to our UI standards.

• It passes all of this to the development agent itself.

• It picks up the ticket in Jira and updates its statuses.

• It runs the tests. • It debates with the review agent to get the code ready.

• It deploys to the test environment.


Today, there are story points within the sprint that AI picks up and completes on its own. For now, it also has a capacity defined in story points; we've capped it. At this stage, we still check everything before it goes live 🙂 But honestly, finding issues is getting harder and harder. For all of this, we use many different tools and methods, such as our BRD AI assistant; in upcoming articles in the series, I'll go through what they do, where and how we run the agents, how we maintain and monitor them, and how we designed security, one by one 🙂


What's next?

There are many topics on our roadmap that we're excited to be working on. Here are a few of them: 


CrewAI: An assistant that talks inside the app. CrewAI knows the documents uploaded from the admin panel. For each document, you can specify which role it applies to, and the assistant is trained accordingly. It can access every piece of data the user can access in the app, so instead of navigating between pages, users can simply ask and learn. Our goal is for users to be able to perform every action they take in the app by talking to CrewAI. How CrewAI behaves for each role and which skills it has will also be configurable dynamically from the admin panel. This will save an incredible amount of time, both in completing tasks and in accessing information. 


EFB tablet inventory module. For the EFB department, we're building a module fed by both Intune and SAP. Tablet inventory can be listed and tracked; entries can be made, and notifications can be created based on many parameters such as version and storage. The rules are dynamic too: "warn me if the version is below this release," "send a notification if free storage drops below this level." Emails and notifications will be triggered automatically. 


Captain and cabin upgrade processes. Alongside the captain upgrade process, we're moving cabin upgrade processes to a system that can be managed end to end, digitally and dynamically. 

E-signed type rating certificates. We had already digitized many forms and trainings; this was the one left. Honestly, we're even a bit late on e-signed, QR-verified type rating certificates, so we aim to deliver them soon. They all share the same goal: moving processes that live on paper, in email, and in people's memory into the crew's pocket.


Closing

It's often believed that big products require big teams. Our experience is a bit different: a well-built platform, a process that genuinely listens to users, and a few simple rules that make decisions cheaper can deliver results far beyond the size of the team. I'm on this journey together with my colleagues on the Crew Solutions team, and I'm proud of every one of them. In the next articles in the series, I'll cover our other cabin and cockpit products, our operational agenda, and how we compare with similar applications. Stay tuned. How does your team manage multiple products at the same time? What was your hardest decision? I'd love to read your thoughts in the comments.


bottom of page