Focus Day

AI for Product Teams

25. november 2026

About the Focus Day

How can product teams use AI to actually build better products?

This focus day explores how AI and agent-driven workflows can support the entire product cycle, not just code generation, but also research, user insights, prioritisation, design, briefs, alignment and reviews.

Through concrete cases and practical experiences from teams already integrating AI into their product development, you will get insight into which workflows and tools create real value, what does not work, and how you can apply the learnings in your own organisation.

The goal is simple: to give you practical inspiration, tools and methods that you can put into use right away.

Who are the speakers?

You will experience speakers from Github, Claude, DX, TV 2 Denmark, Lovable, Ombord.AI and more.

Follow along on this page, and on our LinkedIn where we will publish who the speakers are and what they will be speaking about.

 

Who is it for?

The Focus Day is completely industry-agnostic and extends far beyond the media industry. Anyone working with AI, innovation and product development, regardless of sector, will benefit from the day’s insights, real-world cases and hands-on tools.

Practical information

📅 Date: November 25, 2026
📍Venue: MedieLab, Fyens Stiftstidende, Banegårdspladsen 1, 5000 Odense C

Get Your Ticket
See the program

Speakers

Mathias Bjørn Deichmann

Founder, OMBORD.AI

About the keynote:

TAKE THE PAIN OUT OF PRODUCTION WITH THE SOURCE OF TRUTH

Film production can be tedious. Not because the work itself is complicated, but because information is constantly changing, scattered across documents, spreadsheets, emails and systems. Drawing on years of managing large commercial productions, Mathias will demonstrate a radically simpler approach: one single source of truth, combined with an agent and a tiny deployment platform. The result is less administration, faster changes, greater flexibility — and a way of working that can be applied far beyond film production.

Mathias Bjørn has been a commercial film director and creative director for 20 years. He has always been very interested in tech, and has launched ombord.ai as a platform for developing lightweight and custom AI solutions for the media and communications industry.

aaaaaaaaaa
aaaaaaaaaa
aaaaaaaaaa

Christian Spliid

SYNDICATE

About the Keynote: 

FROM LICENSES TO IMPACT: WHAT WE SEE IN DANISH PRODUCT TEAMS RIGHT NOW

The tools are rolled out, people use them every day, and it is still hard to point to where the work actually got better. We bring the view from the machine room across companies. What do teams really hand over to agents, what stays human no matter how repetitive it is, and why does one person’s AI use hit a ceiling so quickly, while teams that share context and habits keep moving? We look at the two dials that decide the outcome, context and autonomy, and at how the bottleneck has shifted: producing something is no longer the hard part, defining and verifying it is. Leadership expects 5 to 10 times more output, but the measured effect on delivery is 16 to 30%, and only 6% of companies can document a gain across the whole business. This is not an argument against AI. The gain simply comes from redesigning the work, not from handing out licenses.
aaaaaaaaaa
aaaaaaaaaa
aaaaaaaaaa

Bjørnar Almli

Team Lead, TV 2 Norway

About the Keynote: 

AGENT SANDBOX – HOW TV 2 RUNS AI AGENTS IN PRODUCTION WITHOUT LOSING CONTROL

At TV 2 Norway, we have built Agent Sandbox, a platform for running AI agents securely inside our own infrastructure.
Each agent session runs in its own isolated Kubernetes pod, with Entra ID authentication, team-scoped access, network controls and full audit logging.
Agents can interact with internal systems such as our editorial CMS and media archive, while commands, tool calls and output are streamed and persisted.
Usage and cost are tracked per session, user and team, with hard limits that can automatically stop an agent.
The session covers the architecture, security model and networking challenges behind the platform.
It also explores the limitations of strict default-deny networking and the trade-off between guardrails and agent flexibility.
The sandbox enables real production workflows, from advanced research agents to teams orchestrating multiple agents safely.

Victor Foged

Solutions Architect, AWS

About the keynote:

SHIPPING THE CODE NOBODY TYPED

Software development at Amazon is changing shape. Product teams have moved from using AI as a coding assistant to building AI-native: people describe the intent, agents do most of the implementation, and the team spends its time on specifying, reviewing and deciding. Features that used to take a sprint now move from idea to production in a day, and design work that took a week is done in hours. This session looks at that change and at the capabilities that make it possible: agents that can plan, execute and verify multi-step work on their own, codebases and documentation they can navigate, tests they can trust, environments where they can run unattended, and access to the systems a team actually works with. It closes with where this is heading: specifications as the primary artefact, agents working in parallel for hours, and product teams moving closer to the customer as building gets faster and cheaper.

aaaaaaaaaa
aaaaaaaaaa

Amr Al Jarhi

Github

About Github:

Move from AI-assisted coding to AI-powered software delivery.

Discover how GitHub is transforming product development with AI-powered tools like GitHub Copilot and autonomous coding agents. From planning and writing code to testing and reviewing changes, GitHub enables teams to automate development tasks, improve collaboration and focus more on building products that deliver real value

aaaaaaa

aaaa

aaa

a

aaa

aaaaaaa

a

Jess Williams

Client Success Manager, DX

About the keynote:

FROM AI OUTPUT TO PRODUCT IMPACT: WHAT PRODUCT TEAMS CAN LEARN FROM AI-ASSISTED ENGINEERING

AI is making it dramatically easier to generate ideas, specifications, content, and code. But faster output does not automatically create faster delivery or better customer outcomes.

In this keynote, Jess Williams Developer Experience Specialist, draws on current DX research from AI-assisted engineering to explore what happens when AI accelerates one part of the delivery system while bottlenecks remain elsewhere. While the data focuses primarily on engineering, its implications extend across product, design, and technology: clearer intent, stronger shared context, and better collaboration are essential if AI-generated work is to become real product value.

Jess will share practical ways teams can identify where work is slowing down, improve the information humans and AI work from, and apply AI to the constraints that matter most.

Michael Haase

Founding GTM, Lovable AI

About the keynote:

SOFTWARE BUILT BY THE PERSON CLOSEST TO THE PROBLEM

For decades, the ability to create software depended on technical training, capital, or engineering resources. Today, anyone who understands a problem, or is closest to it, can describe it and get working software back.

When domain experts across functions build and run the software their business depends on, security and trust become part of the product: who reviews what gets built, where it runs, and who’s accountable when it touches customer data.

In this talk, Michael shows how enterprises and employees are already bringing Lovable to work, how it’s reshaping what product teams do, and how enterprises can embrace this shift without IT losing a minute of sleep.

Peter Skøtt Pedersen

TV 2 Denmark

About the keynote: 

AI across the product lifecycle: TV 2’s use cases and the work after automation
As agents make code, prototypes and analysis faster to produce, product teams have to spend their time and tokens on building the right thing. That includes understanding users, testing assumptions and deciding what deserves to be built.
I’ll share practical use cases from TV 2 and our product teams, from prototyping features and building internal dashboards to research, analysis and recurring product workflows. I’ll cover what we’ve learned from testing personal and shared knowledge bases, and how domain skills and our Agent Skill and Plugin marketplaces make team knowledge available to agents across the product lifecycle. Code is context and context is code. I’ll show why we need to make our assumptions, evidence and decisions explicit, so we can turn assumptions into experiments and feed what we learn back into shared context. This is the new product work, and it can shorten the distance between intent and value.

Jacob Langvad Nilsson

Fractional CAIO, Co-founder, Applied Futures, og Claude Community Ambassador

When AI Becomes a Colleague: Six New Ways of Working from the Product Teams Leading the Way

In just a few years, AI has evolved from being an assistant handling simple tasks to becoming a colleague capable of working independently for an entire day. This creates new demands on how product teams work. Jacob will show what leading teams are doing differently, and what it looks like in practice. No theory, but real prompts, skills and agents, including the things that did not go as planned.

Jacob Langvad Nilsson is an independent advisor and fractional CAIO (jlangvad.com). He has 15 years of experience with digital transformation and leadership across both public and private organisations, leads the digital and AI area at Mediq, is co-founder of Applied Futures, and serves as an Anthropic Claude Community Ambassador.

Bo Bergstedt

TV 2 Denmark

About the keynote: 

AI is everywhere: in our news feeds, inboxes, images, videos, music and code. It is reshaping education, politics and warfare—and the pace shows little sign of slowing.
This opening talk sets the stage for the day with a snapshot of generative AI in 2026: what it can do, where it falls short, and what its rapid development means for the people building tomorrow’s products.
Bo Bergstedt is Enterprise Architect (AI) at TV 2 Denmark. Over the past few years, he has created AI-generated images, videos and voices for news coverage and entertainment. When he is not making deepfakes, he advises on the opportunities, challenges and practical implications of AI—and builds prototypes to explore what the technology can actually do.