deutsche telekom
An opportunity to position the ready-to-launch AI Phone as a true collaborative partner for users, helping to achieve product-market fit.
Domain
Consumer Technology + Telecommunications
Timeline
Apr‘25 - May’25
STATUS
Proof of Concept (POC)
UNDERTAKEN AT
Deutsche Telekom + Politecnico di Milano
Overview
Deutsche Telekom is a leading European telecommunications provider, renowned for its digital innovation and expansive network services.
In a strategic move to enter the smartphone market, Deutsche Telekom launched an AI phone that reimagines what a mobile device can do by embedding Magenta AI (Deutsche Telekom's proprietary AI) at the core of the experience.
final deliverable
Finding product-market fit by making Magenta AI a true thinking partner, not just delegate to it.
Our primary goal was to find product-market fit for Deutsche Telekom as it entered the AI phone market. We wanted to go beyond creating just another smart device; instead, we focused on designing Magenta AI (Deutsche Telekom's intelligent assistant platform in collaboration with Perplexity) around the concept of "thinking with AI."
This meant making the AI not just a passive assistant or autopilot, but an active thinking partner. Someone users could collaborate with to navigate information overload, make smarter decisions, and bring more meaning and balance into their digital lives.
By emphasizing AI as a cognitive copilot, we aimed to address the fragmented experiences and trust issues that users face with current digital tools.
our driver
Secondary Research
How did we establish a foundation of understanding through secondary research methodologies?
The research methodology combined blue sky research with comprehensive market trend analysis, following established principles of exploratory, curiosity-driven investigation.
Blue sky research, defined as "research without a clear goal" that focuses on fundamental understanding rather than immediate applications, provided the conceptual foundation for exploring AI's role in human cognition.
Blue Sky Research Findings
Literary Analysis
Examined speculative AI narratives from "Snow Crash" by Neal Stephenson and "Atlas of AI" by Kate Crawford, exploring AI as embedded agents and co-pilots.
Product Ecosystem Study
Analyzed existing AI implementations including Google's Smart Reply predictive text, Detroit: Become Human's adaptive responses, and Nest Thermostat's behavioral pattern recognition.
Cognitive Science Integration
Investigated key forms of human thinking (critical, creative, analytical, divergent, logical, practical) and their intersection with AI capabilities.
Signs of anticipation of the future with promising intuitions in movies, exhibitions, arts, games…
Her (2013)
Her is a movie based on an AI that deeply personalizes companionship, blurring boundaries between machine and self. Use the AI for consulting, advisory, partner, friend etc.
Ars Electronica Center
Ars Electronica explicitly explore the boundaries of algorithmic thinking, questioning where human creativity ends and AI begins. It highlights the tension between human agency and machine augmentation-mirroring the archetypal myth of the “helper” or “trickster” that can both empower and disrupt.
Cortana (Halo)
Halo is game that demonstrates a vividly human-like assistant, one that doesn’t just wait for commands, but actively “thinks ahead,” personalizes its support, and becomes a trusted partner in every firefight.
Sunny
Sunny isn’t just a tool; she evolves into a mirror of human emotion. Showrunner Katie Robbins explores this dynamic as “AI as tech, art, or creation, is a reflection back on us.” The series drives home that while AI can bring comfort and companionship, it also holds potential for misuse and manipulation.
Snow Crash: Novel by Neal Stephenson
Stephenson’s fiction Snow Crash imagines personal “agents” that monitor habits and foresee needs.
Insights coming from political, economical, social, technological fields...
Filter bubbles and echo chambers
Ars Electronica explicitly explore the boundaries of algorithmic thinking, questioning where human creativity ends and AI begins. It highlights the tension between human agency and machine augmentation-mirroring the archetypal myth of the “helper” or “trickster” that can both empower and disrupt.
https://www.technologyreview.com/2018/08/22/140661/this-is-what-filter-bubbles-actually-look-like/
Atlas of AI
It is based on Crawford's research into the development and labor behind artificial intelligence, as well as AI's impact on the world.
Algorithmic bias in AI
Rising public concern about algorithmic bias and loss of agency; because biased algorithms can reinforce stereotypes or exclude people, and opaque AI decision-making can make users feel manipulated or powerless. Trust, fairness, and autonomy are at stake. Now and growing-concerns rise as AI becomes more embedded in daily life, especially after high-profile incidents of algorithmic bias or when users notice recommendations that don’t align with their values.
https://www.nature.com/articles/s41599-023-02079-x
Gmail Smart Reply
Gmail’s Smart Reply was introduced for the Gmail mobile apps (Android and iOS) on May 17, 2017; suggests short, context-aware responses to emails so replies can be sent with a quick tap.
Older Americans’ Technology Usage Keeps Climbing
Since the AI operates on natural language commands, it is significantly easier for older generations to learn and use compared to the complexity of navigating different apps.
https://www.aarp.org/pri/topics/technology/internet-media-devices/2019-technology-trends-older-americans/
Principles and constants at the roots of the driver like myths, rituals, legends …
Samara: The wheel of life
Samsara’s cycle of birth, death, and rebirth mirrors AI’s recurring waves of creation, dissolution, and re-formation; each iteration inheriting karma-like consequences from prior designs, data, and norms.
https://hillpublisher.com/ArticleDetails/4132
Delphic oracle
Apollo at Delphi, where a priestess called the Pythia delivered divinely inspired pronouncements to individuals and city-states. Centered in Apollo’s temple on Mount Parnassus, it was the most authoritative Greek oracle from roughly the 8th to 4th centuries BCE, shaping decisions on warfare, colonization, and law.
https://www.britannica.com/topic/Delphic-oracle
Technological singularity
The technological singularity is the theoretical concept that the accelerating growth of technology will one day overwhelm human civilization. Adherents of the idea believe that the rapid advancements in artificial intelligence in the twenty-first century will eventually result in humans either merging with technology or being replaced by it. The concept was first touched upon in the 1950s and later applied to computers in the 1990s.
https://www.ebsco.com/research-starters/computer-science/technological-singularity
Blue sky research
Market Research and Trends Analysis
Generative AI Market Growth
Market size projected to reach $442.07 billion by 2031, representing explosive growth from $7.69 billion in 2021.
AI Usage Typologies
Identified distinct AI categories (Small Language Models, Multimodal, Agentic) serving different user interaction needs.
Business Adoption Drivers
75% of knowledge workers using AI tools, with primary applications in eliminating repetitive tasks and improving efficiency.
What Generative AI Means for Business
Gartner frames generative AI as a transformative business capability that must be aligned to clear use cases, governance, and measurable value, with roles like a chief AI officer emerging to drive strategy and risk management.
AI everywhere: Like magic, but with algorithms
Globally, recreation and entertainment, leisure travel, and restaurants represented an estimated 20% of the consumer’s wallet in August, down slightly from 21% a year ago.
Impact of Generative AI on Critical Thinking
The rise of Generative AI (GenAI) in knowledge workflows raises questions about its impact on critical thinking skills and practices. In other words, when we rely too much on AI to think for us, we get worse at solving problems ourselves when AI fails.
What CIOs Seek From GenAI
CIOs need AI solutions that prioritize productivity and experience gains over near‑term revenue, treating AI as a force-multiplier to streamline work now and monetize later.
Sequoia: System 1 vs System 2 Thinking
This leap from pre-trained instinctual responses (”System 1”) to deeper, deliberate reasoning (“System 2”) is the next frontier for AI. To tackle the most challenging, meaningful problems, AI will need to evolve beyond quick in-sample responses and take its time to come up with the kind of thoughtful reasoning that defines human progress.
ConsumerSignals
Globally, recreation and entertainment, leisure travel, and restaurants represented an estimated 20% of the consumer’s wallet in August, down slightly from 21% a year ago.
Market research and trends analysis
What critical market insights shaped our design direction?
Consumer AI Adoption
60% of consumers consider AI features important when choosing smartphones, with 21% labeling them "very important"
Upgrade Acceleration
AI features driving faster device replacement cycles, with average trade-in age decreasing significantly
Market Polarization
Growing demand for both entry-level AI-enabled devices and premium AI-powered experiences
Primary Research
How did we validate our hypotheses through human-centered design investigation?
The team conducted comprehensive primary research with 20 participants aged between 20 to 60. This research followed human-centered design principles, focusing on empathy-building and need identification.
Investigation covered five critical areas
Cognitive Load & Pain Points in Thinking
Current Strategies for Managing Mental Load
Outsourcing Thinking & Decision-Making
Usage & Trust in AI for Cognitive Tasks
Expectations and Desires for AI Support
What compelling behavioural patterns emerged?
55%
would accept integrated AI if it simplifies life
65%
use AI for research, summaries, or idea generation, but only 15% trust AI for final decisions
Opportunity Area
Which strategic opportunity areas emerged from our research synthesis?
Hybrid Thinking
Supporting combined human-AI cognitive approaches
AI and Cognitive Balance
Balancing between AI assistance and human agency
Evaluating Alternatives
Helping users navigate complex decision spaces
Segmented Thinking & Personalization
Visibility in
Data Usage
Ensuring transparency in AI decision-making processes
Trust and Transparency
Building reliable, explainable AI systems
Natural Interaction
Enabling intuitive human-AI communication
Contextual
Intelligent Memory
Creating AI that understands and remembers context
Key moments with Deutsche Telekom and Politecnico di Milano.
conceptual use cases
We translated the opportunity areas into speculative feature concepts for market-fit.
USE CASE 1
Multitasking Ally
For individuals juggling a demanding blend of work, home, and personal commitments, the AI phone serves as an intelligent, unified cognitive assistant that streamlines fragmented daily tasks.
By proactively taking notes, organizing reminders, facilitating natural delegation, and maintaining contextual memory, the system reduces mental overload and helps users seamlessly transition between roles, creating calm and reliable support for a complex, multitasking lifestyle.
USE CASE 2
Collaborative Decision Space
Empowers families (2-6) in sharing responsibilities to collaborate productively by centralizing coordination, expense tracking, and real-time communication.
With AI-driven conflict detection, group polling, and transparent scheduling, the experience fosters fairness, enabling smoother collective decision-making and reducing misunderstandings in dynamic, multi-user environments.
USE CASE 3
Interconnected Collaborative Canvas
Supports those students and professionals in work, academic, and social spheres by connecting physical and digital workflows into one cohesive, intelligent platform.
Acting as a partner in brainstorming, research, and planning, it bridges handwritten notes and digital thinking, prompts reflective insights, and coordinates group discussions, helping users maintain focus, manage emotional stress, and stay aligned across multiple domains.
future Strategy
What strategic implementation approach would ensure successful market entry?
PHASE 1
Research & Foundations
If we were building this for full-scale implementation, we would spend even more time immersing ourselves in the diverse realities of our potential users.
We’d extend our research beyond initial interviews to include longitudinal ethnographic studies, shadowing individuals as they navigate their daily thinking patterns and tool usage. By bolstering ethical benchmarks and user journey mapping, we’d ensure every insight truly reflected lived experiences and evolving needs.
PHASE 2
UX & Interaction Design
With additional time, we would invest deeply in iterative prototyping and co-design workshops, inviting users to shape not just the interface but also the AI’s personality and mannerisms.
We’d rigorously test conversation flows, decision-making tools, and consensus-building features across varied contexts—refining until interactions felt both natural and trustworthy at every touchpoint. Advanced user testing would let us perfect accessibility, emotional nuance, and cross-device intelligence.
PHASE 3
System & Platform Integration
Given a real-world runway, we’d work to integrate the AI seamlessly with a broad ecosystem, bridging calendars, messaging, notes, smart home devices, and even wearables.
We’d collaborate with data privacy experts to build a robust, user-controlled permission framework, ensuring absolute transparency and trust. Continuous feedback loops and analytics would guide ongoing improvements, cementing the solution as an indispensable daily companion across all facets of life.
Learnings
Key takeaways.
Working with the AI Phone allowed us to wear to multiple hats, it also demonstrates the successful application of comprehensive research methodologies, user-centered design thinking, and strategic market positioning to create innovative AI-powered solutions.