Octopuls - Mena

Discovery opened with a kick-off with the founders, then two remote stakeholder workshops on Miro. Product owners, the CTO, the client's HR consultant, and the developers worked on the same board, so business goals, HR reality, and technical limits surfaced together.

The workshops moved from ambitions to obstacles: how each stakeholder defined the project and its value, who the users were, which features belonged beyond the chatbot, and what could push the product forward or sink it. In a show-and-tell exercise, the developers demoed the existing system, so every idea was weighed against what the AI could really do.

Stakeholder interviews then grounded the product in how HR actually works. A deep dive with the HR consultant covered the five pillars of HR, the ten areas the app would cover, liability, escalation, and how points could be earned at no cost to the employer.

Desk research and a competitive analysis, from HR chatbots such as inFeedo's Amber and AskHR to engagement platforms such as Officevibe, Culture Amp and Qualtrics, showed what the market already did well and where an AI-plus-human tool for Arabic-speaking teams could stand apart.

About the Research

Our research combined stakeholder workshops and interviews with desk research and a competitive analysis, covering HR end-to-end, from an employee's first question to the moment a human steps in:

• Workshop · Project debriefing and USP: each stakeholder defined the project and its unique value in their own words.
• Workshop · Sailboat exercise: opportunities, obstacles, risks, and objectives, mapped as wind, anchors, rocks, and an island.
• Workshop · Users, needs and concerns: eleven user types, prioritised and grouped into app users and admins.
• Workshop · Feature listing: every feature stakeholders wanted, from points and rewards to knowledge.
• Workshop · Developer show-and-tell: a demo of the existing system, its AI, and its constraints.
• Stakeholder interviews: a deep dive into HR pillars, the ten HR areas, liability, escalation, and gamification.
• Competitive analysis: 22 HR chatbots, sentiment-aware bots, and engagement platforms, compared feature by feature.
• Desk research: company sizes, HR structures, survey design, and the trade-offs of anonymity.
• Feature and user story review: feature lists, personas, and user stories reviewed with stakeholders.

The research converged on one reframe: the bot was never the product. Octopuls had to be a listening system, in which every chat, check-in, and survey feeds the same picture of how people feel, read through sentiment analysis of mood check-ins and conversations. Trust had to be designed in, not promised in the terms. Nine insights from the workshops, the HR interview, and the market review set the rules that every later decision was checked against, from what the AI may remember to how rewards are paid for.

Octopuls serves two sides of one relationship. The user-listing exercise started wide, with eleven user types, from interns and new hires to managers, IT support, and compliance officers. The stakeholders narrowed them to two groups: the employees who use the app, and the HR and admin teams who act on what it learns. Three personas came out of the user stories: the employee, the newcomer, and the manager. The HR team is shown for context. Mindflares designed the employee app; the admin dashboard remained the client's concept, planned for a later phase.

With the rules defined, ideation became a question of habit and trust: how do you get people to open an HR app every day, and then tell it the truth? We turned the insights into six design principles. They decided what made the MVP and how every screen should behave. Rewards, for example, had to cost the employer nothing, so points, streaks, and redeemable rewards were chosen over local partnerships.

Every feature needed a home in a structure simple enough for daily use. Early drafts gave Schedule its own tab; as the flows matured, Knowledge took its place. The result is five pillars, with the AI assistant at the centre of the bottom navigation:

• Home: customizable widgets for mood, announcements, calendar, points, quick actions, and survey pop-ups.
• Chat: anonymous and named conversations, by text or voice, with suggested and action-based prompts.
• Surveys: pulse surveys to answer now or later, with results for knowledge quizzes.
• Knowledge: company material, news, FAQs, and learning content by theme.
• Profile: rewards and ranking, mood trends, calendar, and personal details.
   

User Flows

We turned the user stories into eight goal-based flows. Each one starts from what an employee is trying to do, not from which screen they are on. Two carry the core story of early detection.

The final interface turns HR from a help desk into a daily conversation. A calm, dark UI keeps the focus on the few things that matter each day: how you feel, what you need to know and who can help. Visually, Octopuls lives in deep navy with faint wave-line textures, lit by its blue octopus mark and cyan and green accents. Proxima Nova keeps the tone friendly and legible, and every screen ships in dark and light modes.

Final UI Design: Key screens