AI-LEAD · Path 2: Lead and create value
Leading in the Human-Agent Organization
Two days after which your managers know what to delegate to an assistant and what never to.
- Duration
- 2 days
- Group
- up to 12 people
- Format
- In-house · Remote
- Language
- German or English
- Price
- €15,800 flat, plus VAT
Who this track is for
Managers with people responsibility, from team lead to department head, up to twelve people. For companies where assistants are already available or about to be rolled out while the leadership level has not yet settled how to handle them.
Who it is not for
Not for the executive team, which first needs the fundamental decision and starts in the Sovereign AI Executive Briefing (AI-EXEC). Also not for staff without a leadership role, who belong in AI-WORK.
Starting situation
The tools are there or on their way, but the leadership questions are open: Which tasks may go to an assistant, who is accountable for its drafts, how do you talk to a team that fears surveillance or replacement? Some managers delegate everything and review nothing, others block everything. Both cost trust and results.
This exists after the track
- A delegation matrix per manager: which task classes go to assistants, which to agents with approval, which stay with humans
- A decision and escalation model with approval levels built on the principle: the assistant drafts, the human approves
- A 90-day adoption plan for the own team with the first three leadership routines
- A robust line of conversation for works council and team that takes fears seriously instead of talking them away
- A documented role profile of the manager in the human-agent organization
Prerequisites
People responsibility and some first-hand contact with an AI assistant. Anyone who has never worked with an assistant catches up in one hour with the prework.
Preparation before the track
Each manager breaks a typical working week into ten recurring tasks in advance and notes an assessment for each: would I delegate this, to whom, and what holds me back. These lists are the working material of day 1.
What's included
- Two on-site or remote days delivered by Dino Bordonaro
- Templates for delegation matrix, escalation model and adoption plan for further use in your company
- Conversation guide for team and works council discussions
- Certificates and audit-proof training documentation as a building block of your Article 4 evidence
- 60-minute remote group consultation four weeks after the training
Agenda
Day 1
Your leadership week under the X-ray
Work on the week lists brought along: Which of the ten tasks are drafting work, which are genuine decisions? First sorting on the wall, and the patterns across all participants become visible.
What assistants and agents genuinely deliver today
Live work on the system: drafting a target agreement, preparing a difficult feedback conversation, summarizing a meeting. Including one visible error and the question: How would I have caught it if I had only skimmed?
Delegation rules
Building the task-class matrix: goes to the assistant, goes to the agent with approval, stays with the human. Exercise on the own week lists, edge cases are decided by the group. Guiding principle: the assistant drafts, the human approves.
Decision and escalation model
Defining approval levels: When is a spot check enough, when does it take four eyes, when is AI off limits? Three critical cases are played through, including a flawed draft that almost went out.
Day result: your delegation matrix stands
Each manager documents matrix and approval levels for their own team and presents them to a partner in a peer review. Open points are noted for day 2.
Day 2
Taking fears seriously
The real objections on the table: Will this replace me? Is my work being measured now? Exercise: formulating answers that are honest instead of soothing, and naming what a manager can promise and what not.
Works council and co-determination
Which rollout steps can be subject to co-determination, what early involvement looks like in practice and which commitments hold. No legal advice, but a conversation structure that holds up in the meeting with the council. Exercise: opening the works council conversation in a role play.
The adoption plan
The first 90 days in the own team: the manager's visible own usage, three team routines, measuring success without a feeling of surveillance. Each manager decides: Which task does my team start with in week 1?
Leadership routines live
Simulation of a weekly with assistant drafts: review the draft, grant or refuse approval, formulate feedback to the team. The approval rules from day 1 are applied under time pressure.
Day result: your plan in front of the group
Each manager presents delegation matrix and adoption plan in five minutes while the group pushes back with the hardest objections. The fixed version goes into the training documentation.
Exercises and lab share
Around half of the time is exercise: live work with an assistant on the participants' own anonymized leadership tasks, role plays for team and works council conversations, and a weekly simulation with real approval decisions.
Platforms
Work takes place in your own assistant tenant, for example Microsoft 365 Copilot, or in our lab environment with the Sovereign Assistant, each in the version available at the time. Confidential personnel topics are handled in anonymized form only.
Transfer evidence
The transfer evidence is the presentation of delegation matrix and adoption plan in front of the group at the end of day 2. Both artifacts and attendance are documented in an audit-proof way.
Artifacts you take home
- Delegation matrix for the own team with approval levels
- Decision and escalation model as a document
- 90-day adoption plan with three leadership routines
- Conversation guide for team and works council
- Role profile of the manager in the human-agent organization
Optional extensions
- Use Case and Business Case Sprint (AI-VALUE) when the team ideas should become a portfolio
- AI Champions Program (AI-CHAMP) for multipliers from the participants' teams
- EU AI Act and AI Governance (AI-GOV) when policies and oversight questions come up
Boundaries
The training enables managers. It replaces neither employment-law advice, nor a works agreement, nor the rollout of the tools themselves. For policies and governance structures, AI-GOV is the fitting track.
Frequently asked questions
Can our works council take part?
Yes, and in many companies that is a smart move. One or two seats for works council members noticeably change the later rollout discussion, because both sides work with the same vocabulary and the same escalation model.
Does the format work remotely?
Yes, both days run fully remote, including the role plays in breakout rooms. In our experience the fear and works council topics are denser in person, so we recommend at least day 2 on site.
Our managers have very different levels of prior knowledge. Is that a problem?
No. The prework brings everyone to a common baseline, and the work happens on each participant's own task lists, not on someone else's examples. Those who already delegate a lot sharpen their approval rules, and the skeptics work on defensible boundaries.