Overview

Blinc was our bachelor's degree project at Malmö University, developed together with Mikaela Rasmusson. The project had two outputs: a qualitative research study exploring how generative AI reshapes the professional identity and working methods of project managers, and a functional UX prototype designed as a direct response to what the research found.

The research was anchored in Erving Goffman's dramaturgical theory — the idea that professional life is a kind of performance, divided into frontstage (what clients and stakeholders see) and backstage (the preparation, creative work, and collaboration behind the curtain). With generative AI entering that backstage, we set out to understand what it does to the person performing.

"The question isn't whether AI changes what project managers do. It's whether it changes who they feel they are."

The Problem

Generative AI is moving into professional workflows at pace — but the conversation has been almost entirely practical. Productivity gains, prompt engineering, which tools work best. What we found almost entirely absent from both industry discourse and academic literature was the human dimension: what happens to professional identity when AI starts doing part of your job?

For project managers, the tension is especially sharp. The role is built on credibility — the ability to understand complexity, communicate it to people who don't have time to read a 40-page report, and be trusted when you say the project is on track. When a language model can draft that report and generate the slides in under two minutes, the question of where expertise lives becomes real.

Research gap

• Efficiency, not identity — existing studies focused on productivity metrics, not how practitioners felt about AI-generated work
• Professional pride unexplored — no qualitative research had examined what happens to a professional's sense of ownership when AI is involved
• Goffman not applied to AI — dramaturgical theory had never been used as a lens for understanding AI's impact on knowledge workers

Research

We conducted eight semi-structured interviews with project managers and team leads based in Malmö and Helsingborg. Participants were recruited through LinkedIn and professional networks, spanning industries including tech, construction, marketing, and management consulting. Interviews ran between 45 and 75 minutes.

Theoretical framework

Goffman's dramaturgical model divided the analysis. Frontstage describes moments visible to the audience: client pitches, stakeholder presentations, project reviews. Backstage covers preparation — team ideation, drafting, internal alignment, and everything that enables the frontstage performance without being visible in it.

This distinction proved immediately generative. As soon as we introduced the frontstage/backstage language in interviews, participants naturally mapped their own AI use onto it — and the pattern was consistent: AI belonged backstage.

Interview structure

Each interview moved through four sections: (1) background and current role, (2) creative processes and daily workflows, (3) AI use and personal working methods, (4) views on the future of the profession. The sequence was deliberate — establishing context and rapport before asking about authenticity and professional pride.

Thematic analysis

Analysis followed an inductive approach — themes were not defined in advance but allowed to emerge from the material. Transcripts were color-coded to tag recurring concepts. Physical mind maps were built on whiteboards to trace relationships across participant responses. Two full iterations of theme development were completed and debated before the final framework was agreed.

Findings

Four major themes emerged from the interviews. Taken together, they paint a more nuanced picture of professional identity in the age of AI than we anticipated.

AI as backstage sounding board

Every participant used generative AI — but almost exclusively backstage. It was a brainstorming partner, a document structurer, a first-draft generator for internal use. No one brought it into a client meeting. The frontstage performance remained entirely human — AI shaped it, but was never visible in it.

Authenticity tied to control, not authorship

We expected authenticity to be about whether you wrote the words yourself. It wasn't. What mattered to participants was control: did I decide what went in? Did I set the direction, review the output, make the judgment calls? If yes, the work felt authentic — regardless of whether a model produced the first draft.

Professional pride intact — and potentially enhanced

Pride in one's work was not diminished by AI involvement. Several participants described the opposite: when AI handled the mechanical parts, they had more space to focus on what they considered genuinely skilled — strategy, client relationships, and communication quality. The performance felt more considered.

Role shifting toward facilitation

A subtle but significant pattern: participants described themselves less as producers of content and more as curators and quality assurers of AI output. The role was shifting from generation to direction — knowing what good looks like, and steering AI toward it.

The Prototype

If project managers are already using AI backstage to build presentations — and our research confirmed they are — the design question becomes: what would a purpose-built tool look like? One that keeps human judgment central, makes the AI's role transparent, and lets the PM feel like an art director rather than a passenger.

Blinc is that tool. An AI-powered pitch creation platform designed specifically for project managers and presenters, built on everything the research taught us about how practitioners actually work.

Dashboard

The home view surfaces what normally stays invisible: 12 active projects, 4 pitches generated per week, 37 hours saved per month, 218 generated elements. These metrics aren't decoration — they make the value of AI-assisted work concrete, which participants in our research consistently undersold when describing their own productivity.

Creating a pitch

The creation flow starts with a natural language description — the PM describes the pitch in their own words, as if briefing a colleague. From there, every meaningful decision stays human:

• Audience — Stakeholder / Investor / Board of Directors / Internal team / Client
• Tone — Trustworthy / Visionary / Analytical / Warm
• Length — 6 / 9 / 12 / 18 slides
• Background material — upload documents, data exports, or context files

This setup was deliberate and direct from our research. Every creative decision — audience, tone, structure, intent — remains with the PM. The AI executes; the PM directs.

The generation process

Rather than a spinner, Blinc shows exactly what it's doing: parsing the prompt, building narrative structure (Hook → Problem → Insight → Solution → Ask), sketching visual direction, generating slide titles, composing elements, finalizing the draft. Transparency here was a direct response to our finding that trust in AI output is closely tied to understanding how it arrived at its answers.

Slide editor

The final editor is where the PM takes full control: adjust layouts, switch themes, add animations, chat with the AI assistant for targeted revisions, or generate ambient audio for the presentation. The goal was to make editing feel like creative direction — not damage control on something you didn't make.

Learnings

This was the most research-intensive project I've worked on — and the most personally resonant. Eight conversations about professional pride, authenticity, and what it means to be good at something when a machine is getting better at it too. That changed how I think about designing for AI.

The core lesson: technology rarely changes what people fundamentally want. The project managers we spoke to still wanted to feel capable, trusted, and in control of their work. A well-designed AI tool doesn't threaten that — it amplifies it. The failure mode is building something that makes the person feel like a passenger. The success mode is building something that makes them feel more skilled.

Practically: working across qualitative research and UX design in the same project clarified something I'd suspected — the most useful design decisions come from understanding what people are trying to protect, not just what they're trying to do.