3 tracks · 12 categories

The tracks of AI Product Day 2027

The program is organized in three tracks and twelve categories. Pick one primary category for your proposal. If your talk could fit in two categories, the placement rule and the boundary cases below will help you choose.

How to place your talk

Ask yourself what your talk is mainly about:

  1. What you build for your customers, and how you learn what to build Track 1
  2. How you build Track 2
  3. How the company organizes, equips and governs itself Track 3

Choose the category that matches the main contribution of your talk, even if the talk touches other topics.

Track 1

AI Products & Experiences

Customer-Facing Products, Interfaces & Agent Users

This track covers the products you put in your customers' hands, for human users and for the agents that act on their behalf. Products built for your own employees belong in Track 3.

  1. 01 Building AI-native products, from discovery to scale

    We want the inside story of products where AI is the core: how the team discovered what to build, which bets it made, and what happened between the first version and scale.

    Topics we expect

    • Discovery with AI: augmented user research, synthetic users and their limits, prototyping to learn
    • Product bets: when AI is justified and when a simple workflow, or nothing at all, works better
    • Defensibility when the model labs can ship the same feature next quarter
    • The full journey of an AI-native product, from first version to scale

    We want to hear from

    Founders, product leaders, PMs and UX researchers who shipped an AI-native product and can tell the whole story, including the dead ends.

  2. 02 Designing experiences for humans and agents

    Users now talk to products, delegate tasks to them, and sometimes send an agent in their place. This category covers design for human users and for agent users.

    Topics we expect

    • Generative, conversational and voice interfaces in production
    • Autonomy levels: what users let an agent do and how they take back control
    • Explaining AI behaviour and recovering from errors without losing user trust
    • Agents as users and buyers: agentic commerce, GEO, how machines find and choose a product

    We want to hear from

    Designers, PMs and engineers who redesigned an experience around AI or around agent users, with evidence of how real users responded.

  3. 03 Evals & context engineering: who defines quality?

    The quality of an AI product depends on who defines "good enough" and on the context its model receives. Talks can cover either side.

    Topics we expect

    • Evals written by domain experts, not only by engineers
    • Monitoring and correcting AI behaviour in production
    • Context engineering for customer-facing agents: what the agent needs to know and how you keep that knowledge current
    • The quality trade-offs you made and how you measured them

    We want to hear from

    PMs, domain experts, AI engineers and data practitioners who own the quality of an AI product in production.

  4. 04 Value, adoption & economics of AI features

    This category covers the numbers that a demo does not show: usage, retention, costs and margins.

    Topics we expect

    • Real usage of AI features compared with what the demo promised
    • Why users adopt or abandon an AI feature
    • Inference cost, cost per useful outcome and their effect on the roadmap
    • Pricing and margins for AI products

    We want to hear from

    PMs, product leaders, founders and growth or finance partners who can share measured results, good or bad.

Track 2

Product & Tech: How We Build Now

Builder Practices, Delivery Models, AI-SDLC

This track covers how product, design, engineering and data people work with AI every day, from one person's setup to the way a delivery team ships and maintains software.

  1. 05 Product Builders: working across crafts

    Designers now ship code, engineers take on product work and PMs prototype in production. This category covers what changes in the work when one person crosses these borders.

    Topics we expect

    • The tasks you now handle that were not in your job description a year ago
    • Design systems as agent context, and the move from mockup to code
    • What a team gains and loses when one person covers several crafts
    • How review and quality hold up when more people ship

    We want to hear from

    Designers, PMs, engineers and data practitioners who changed how they work and can describe the situation before and after.

  2. 06 Second Brain & the builder's workstation

    Show us the tools, memory and routines you use with your agents every day. We prefer a live demo or detailed screenshots to a description.

    Topics we expect

    • Personal setups with agents: tools, configuration, routines
    • Second Brain and personal knowledge systems that agents can use
    • What improved your daily work and what turned out to be a distraction
    • The maintenance cost of a personal AI setup

    We want to hear from

    Builders of every profession who use their setup daily and can demo it live or show it in detail.

  3. 07 Delivery models: who owns the result when agents do part of the work?

    When agents take on part of the work, the delivery team changes too. This category covers the team's size, roles and rituals.

    Topics we expect

    • Team size and composition with AI in the loop
    • Roles, handoffs and ownership of the outcome
    • Rituals that changed, disappeared or appeared
    • Moving from individual productivity gains to team performance

    We want to hear from

    Engineering managers, product leads, tech leads and team members who reorganized how their team delivers.

    Decisions that span several teams (portfolio, funding, headcount) belong in category 12. Category 12: Operating models & AI governance

  4. 08 The AI-SDLC, from spec to maintained code

    AI changes every step of the software development lifecycle. Many teams find that the hardest problems start once the prototype works.

    Topics we expect

    • Specs as the interface with coding agents
    • Orchestrating coding agents in a real codebase
    • Code review as the new bottleneck
    • Testing, securing and maintaining AI-written code over time

    We want to hear from

    Engineers, tech leads, platform teams and engineering leaders who run an AI-assisted development lifecycle beyond the pilot and know what it costs to maintain.

Track 3

The Agentic Company

Agentic Workspaces, Platforms & Organizations

This track covers how the whole company organizes, equips and governs itself with AI, including product and tech organizations and every business function.

  1. 09 Agentic platforms & Agent Builders

    Companies build shared platforms so that employees can use, build and share agents, and they create new roles to do that building.

    Topics we expect

    • Internal agentic platforms: build or buy, shared capabilities, controls, adoption by employees
    • Internal AI products built for one team or one function
    • Agent Builders inside business teams: who they are and how they work
    • Forward Deployed Engineers: where they sit in the organization, what they own and how they work with business teams
    • Security, sovereignty, supplier dependence and multi-model choices for these platforms

    We want to hear from

    Platform leads, CIOs, CTOs, security leads, Agent Builders, Forward Deployed Engineers and business teams who built agents for their own company.

  2. 10 Company Brain: what should agents know about the company?

    Agents can only use the company knowledge they can reach. Building that shared memory raises questions of ownership, access and freshness.

    Topics we expect

    • Company knowledge and memory for agents: what goes in and who owns it
    • Access rights, freshness and conflicting sources
    • Correcting what the agents know when it is wrong
    • Partial implementations and failures, as well as mature cases

    We want to hear from

    Knowledge, data and platform teams, and business leaders who built or tried to build a shared knowledge layer for agents.

  3. 11 People management & upskilling

    AI changes what people do every day and what their managers expect of them. This category covers how companies help people build new skills and how management changes along the way.

    Topics we expect

    • AI fluency and upskilling programs: what people actually learned and how you measured it
    • Managers as adoption levers: using AI visibly, setting expectations, evaluating work done with AI
    • Juniors, careers and hiring when entry-level tasks change
    • Cognitive load, autonomy and how people experience working alongside agents

    We want to hear from

    HR and learning leaders, managers and employees who changed how people learn, work and are managed, and can show what happened.

  4. 12 Operating models & AI governance

    Local gains from AI do not always reach the P&L, and when an agent acts, someone in the company has to answer for it. This category covers decision rights: how the company redesigns its work and allocates its AI investments, and how it governs agents without blocking the work.

    Topics we expect

    • Business workflows redesigned by the people who do the work, in support, sales, operations, finance or HR, and whether the gains show up in business results
    • Authority across teams: portfolio priorities, funding, headcount
    • Accountability for decisions made or prepared by agents
    • Shadow AI and the AI Act in practice: what you found and which requirements apply to you

    We want to hear from

    Executives, transformation leaders, CISOs, legal and compliance practitioners who changed how their company decides about AI and governs it, and can show the results.

Boundary cases

Found your category?

The call for proposals is open until January 3, 2027.