Designing your AI Twin: highlights from a Helsinki masterclass

On 26 November 2025, the Hands-On AI Masterclasses brought together practitioners, innovators, and CX leaders for a full day of interactive learning at the Scandic Marina Conference Center in Helsinki. Among the practical workshops, one session stood out for its direct relevance to the future of personal AI ownership: Design Your AI Twin: Human-Centric Intelligence in Action, led by Valto Loikkanen (Prifina), Sirte Pihlaja (Shirute), and Markku Pelkonen.

Along with their own AI Twins as well; Valto AI, Sirte AI and Markku AI.

What follows is an in-depth recap of the session for everyone exploring how personal AI twins are created, how they operate, and why this shift matters now.

Why Personal AI Twins Matter

The session opened with an important reminder: individuals play many roles throughout their lives. We move between professional responsibilities, customer perspectives, and personal identities. As AI transitions from simple tools into intelligent agents, individuals need an equivalent digital presence operating fully on their side.

For most mainstream AI assistants today, primary incentives are tied to the platform or the investor. This means the assistant cannot simultaneously act in the best interest of every user. A personal AI twin, in contrast, exists solely for the individual. It is designed around user-owned data, user-defined boundaries, and user-controlled knowledge.

This distinction becomes critical as the world moves from isolated AI apps toward multi-agent ecosystems, where both customer-side and service-side systems are interacting autonomously. In that world, you need an agent that negotiates, filters, interprets, and interacts on your behalf.

Prifina’s Model: Information Ownership as the Foundation

Prifina’s approach is built on a clear principle: your information should remain under your ownership and control.

Valto Loikkanen outlined several defining components of this model:

• ·Your data stays with you and is not used to train foundation models.
• ·The model and your personal information are kept separate, with the model simply acting as an interface.
• ·Your AI twin is portable. In the future, just as you can move your phone number between operators, you should be able to move your personal information and AI twin between service providers.
• ·No tracking or surveillance. Public-facing AI twins are not indexed by search engines and do not track visitors.

!2025_11_26_cxmasterclass_Web-299.jpgThis ensures that personal AI is not tied to a single dominant platform but can exist in a healthy, competitive environment where individuals maintain agency and choice.

Inside the Workshop: Building an AI Twin in Real Time

Participants were invited to create their own AI twins right in the session. The process included:

• ·Email verification to establish ownership.
• ·Naming the twin, which determines the public-facing identity.
• ·Accessing the AI Twin Ownership Portal, the control center for customization.
• ·Creating a profile: description, image, disclaimers, and external links.
• ·Configuring behavior: interaction style, tone, response length.
• ·Adding knowledge through text, copy-paste inputs, or uploaded files.
• ·Reviewing interactions through a built-in conversation log that shows how the twin answers and how well it matches available knowledge.

All knowledge can be added in multiple languages, with the twin adapting responses automatically.

The guidance emphasized using publicly shareable information only, allowing participants to get comfortable with the system while avoiding sensitive personal data.

Other team members, supported participants hands-on throughout the creation process.

Real-World Examples: Diplomats, Trainers, and Talking Products

Valto showcased several live examples of AI twins in use:

• ·A diplomat using an AI twin trained on an autobiography and case materials to support media interviews and stakeholder conversations.
• ·A lawyer-oriented twin that screens and prepares client questions before meetings.
• ·Talking Product and Talking Venue AI powered by the same underlying technology, enabling physical products and locations, such as books, wine bottles, gardens, and the Mall of Tripla, to speak directly with consumers and gather insights.

!2025_11_26_cxmasterclass_Web-296.jpgThese illustrate the wide range of cases where user-owned data and programmable personal agents create new types of interaction and value.

Markku Pelkonen: Living and Working with an AI Twin

The latter part of the session focused on practical long-term use of an AI twin, presented by Markku Pelkonen, whose expertise spans education, leadership development, and organizational transformation.

!2025_11_26_cxmasterclass_Web-313.jpgMarkku uses his AI twin daily to scale his presence across multiple countries and projects. His twin holds:

• ·Hundreds of pages of documentation
• ·Dozens of hours of training videos
• ·Structured models and rubrics used in educational programs

During a three-day training program, participants could ask either Markku or his AI twin questions. Many preferred the twin’s ability to deliver clear, concise, and contextually consistent answers.

For Markku, the AI twin has become:

• ·A colleague
• ·A scalable extension of his expertise
• ·A way to remain present across simultaneous engagements
• ·A living archive of his thinking, frameworks, and problem-solving methods

He trains his AI twin continuously, often by recording thoughts during walks, transcribing them, and adding them to the knowledge base.

The Skills That Matter Most in the Age of AI

When asked about which skills are most important as AI becomes ubiquitous, Markku highlighted:

• ·Adaptability
• ·Resilience
• ·Creativity
• ·Curiosity
• ·Learning to learn
• ·Growth mindset
• ·Systemic thinking
• ·Emotional and social skills

He emphasized that AI literacy is not just about prompts; it is about context, interpretation, and the ability to understand complex systems.

A Beginning, Not an End

The session closed with a simple reminder:

Your AI twin will never be finished.

It is an evolving digital counterpart that grows with you as you add information, refine its purpose, and expand its abilities.

For many participants, the session was the first step in understanding the personal value of their information and the role their AI twin can play in the coming agentic future. Prifina’s view is that personal AI will soon become as essential as having a phone number or email address. Starting early makes all the difference.

If you would like to explore personal AI twins yourself or integrate them into your work or community, Prifina’s platform is open, with new capabilities becoming available continuously.

For more information or to get started, visit prifina.com.