01 · Executive Seminar  ·  Agenda & Format

One day. Six artifacts. The conversation your leadership team probably hasn't had yet.

A structured working session for leadership teams who are already building agentic systems — and need to step back and decide what they're actually creating before they keep going.

Format
1-day working session
Participants
Up to 15 leaders
Duration
7 hours, 4 blocks
Output
6 working artifacts

Most leadership teams are already building agentic systems. They just haven't decided what they are.

The Executive Seminar isn't a workshop and it isn't training. It's a structured one-day working session that gets your leadership team out of fragmented decision-making and into a single, shared model of what you're building, what's at stake, and what has to become explicit before you keep going.

By the end of the day, your team has named the agents you're already creating, mapped where you're exposed, and produced six concrete artifacts you can act on Monday morning.

Four working blocks. Each ends with a decision, not a slide.

Block 01
09:00 — 10:30
90 minutes

Where you actually are

"What kind of behaviour have we already introduced into our product?"

Most teams discover during this block that they're further along than they thought — and less aligned than they assumed. We map your current AI surface area honestly: what's shipped, what's behaving inconsistently, what was decided implicitly, and where alignment is assumed but not real.

  • Walk current AI/agent surfaces — shipped, in build, in consideration
  • Place each on the AUX Evolution Curve (tool → assistant → agent → autonomous)
  • Identify the decisions that were made by default, not by design
  • Surface the gap between the team's individual mental models
Block 02
10:45 — 12:45
120 minutes

The four behaviours you're introducing

"Where are we exposed — and which exposures are we still postponing?"

Agentic systems introduce four behaviours classical software doesn't have: interpretation (intent), continuity (memory), uncertainty (confidence), and action (autonomy). Each requires an explicit model. We work through each, naming the decisions your team has already made, the ones you've postponed, and the risks created by the postponement.

  • Walk the AUX trust ladder — functional → contextual → judgment → advocacy
  • Map your current systems against intent, memory, confidence, and autonomy
  • Identify the trust gaps your users, customers, or auditors will notice first
  • Decide which gaps are acceptable and which are existential
Block 03
13:45 — 15:45
120 minutes

Trust design intensive

"For one specific system — what should it actually be allowed to do?"

The day's most concentrated working block. Pick one of the agentic systems surfaced in Block 01 — usually the one with the highest stakes or the most ambiguity — and work through its full design contract live, with the leadership team contributing context. This is the conversation that almost never happens between a CEO, a CTO, and a head of product simultaneously.

  • Map the autonomy contract — what's automatic, what's confirmed, what's blocked
  • Design the trust signals — confidence cues, escape hatches, intent handshakes
  • Define the memory policy — what it remembers, forgets, and asks consent for
  • Run the procurement / governance test — would this survive review?
Block 04
16:00 — 17:30
90 minutes

What changes Monday

"Which decisions can no longer be deferred — and who owns them?"

The closing block converts the day's work into specific commitments. We name the decisions that have to become explicit, assign owners, and produce a 90-day plan that ties to your existing roadmap rather than running parallel to it. The goal is not "more things to do" — it's removing the ambiguity that's already costing you.

  • Decisions that move from implicit to explicit, with named owners
  • What stops, what starts, what continues
  • 30 / 60 / 90-day commitments tied to existing roadmap
  • Re-engagement criteria — when to bring auxfirst back, and when not to

Six working artifacts your team can use on Monday.

The seminar produces documents, not slides. Each artifact is a deliverable your team made together — which is what makes them stick. They become reference material in subsequent product, engineering, and stakeholder conversations.

A1

Current State Map

Your AI/agent surface area placed honestly on the AUX Evolution Curve, with notes on what's shipped, in build, and in consideration.

A2

Trust Gap Register

Where intent, memory, confidence, and autonomy are under-designed in your current systems — ranked by exposure.

A3

Autonomy Contract

For the system worked on in Block 03 — what's automatic, what requires approval, what's blocked, and the rationale for each line.

A4

Trust Signal Specification

Confidence cues, intent handshakes, escape hatches, and memory policy — designed for the same system, ready to hand to product and engineering.

A5

Decision Register

Decisions that moved from implicit to explicit during the day, with named owners and review dates. The most-referenced artifact post-seminar.

A6

30 / 60 / 90 Plan

What changes in the next quarter — phased and scoped to your existing roadmap, not added to it.

The day works because of what wraps around it.

The session itself is one day. But the engagement runs from briefing call to follow-up — about three weeks end-to-end. The structure around the day is what makes the day work.

Before · The week prior

Preparation

  • 30-minute briefing call with the engagement sponsor (CEO, CPO, or CTO)
  • You share: current product surfaces, internal AI tool inventory, recent customer-facing decisions
  • auxfirst sends: the AUX framework primer pack for all attendees, custom prep brief tailored to your context
  • 2–3 short interviews with named participants, where context warrants
After · The 30 days that follow

Follow-through

  • All six artifacts delivered as polished, branded leave-behinds within one week
  • Recorded session walkthrough for non-attending stakeholders, where requested
  • 30-day follow-up call to check progress on the Decision Register and the 90-day plan
  • Optional bridge to Blueprint Sprint or Agent Validation Sprint if a specific build is greenlit

Built for leadership teams. Not for individual contributors.

The seminar produces value because it gets people who normally don't sit in the same room sitting in the same room. We strongly recommend at least two of the three roles below — and ideally all three.

Role 01

Strategy & commercial

CEO, founder, managing partner, or commercial lead. Owns the question of what we should build and what we should promise the market.

Role 02

Product & design

CPO, head of product, head of AI, or design leadership. Owns the question of what the system actually does and how users experience it.

Role 03

Engineering & risk

CTO, VP of engineering, head of platform, or — for regulated contexts — legal/compliance. Owns the question of what's safe to ship and what we'll be accountable for.

The biggest risk isn't building agents. It's building them without realising what they are becoming.

Send us your situation.
We'll tell you if this is the right room.

A 30-minute scoping call to understand where you are, who should be in the seminar, and whether the Executive Seminar is the right next step — or whether something else fits better. No obligation, no deck.

Book a scoping call

Tell us where you are and who would be in the room. We'll come back with a view on fit within one business day.

View All Engagements
Emil Krzemiński · auxfirst