Journey Steps
Every step the participant takes in the Empower Companion, grouped by session. Each step has a done when check, so participants know when they can move on, and a catch-up shortcut for when they fall behind.
Note
Step IDs (J1, J2, ...) are used by trainers to set the current step for the group. They are a proposal and may change after the dry run.
Session 2 – Choose
Happens during Part D – Backend Handover.
| Step | Participant does | Companion provides | Done when | Catch-up |
|---|---|---|---|---|
| J1 Welcome and profile | Enters a display name, optionally a company and email address, and accepts the data notice | The goal of the journey, the session overview, and what is shared at the end | A profile exists | – |
| J2 Choose vertical | Chooses one of the verticals deployed on their system | A short description per vertical, its top-level object, and its terminology (see Vertical Terminology) | A vertical is selected | The trainer's suggested vertical for the event |
| J3 Choose problems | Chooses one or more problems and marks which ones to build now: one for most participants, all three for advanced participants | The problem cards: user role, key question, and context for each of the three problems of the vertical, and the levels | At least one problem is selected and marked to build now | Pick the problem the trainer demonstrates |
| J4 Check backend | Runs the backend check | A check that the vertical's DOM module, services, and elements exist and that the scenario is running | All checks are green, or the participant has asked a trainer for help | Trainer note with what to fix |
Session 3 – Define and Build
Follows the App Builder Workshop session flow. Steps J6 to J8 replace the paper Outcome Canvas.
| Step | Participant does | Companion provides | Done when | Catch-up |
|---|---|---|---|---|
| J5 Explore the data | Opens the data model of their vertical and looks at live alarms | The data model cheat sheet and links to the vertical's view and DOM module | The participant has checked off the objects and data their problem needs | – |
| J6 Define the end state | Describes the ambitious future state the chosen problems contribute to | Guiding questions: Who benefits? What is different? What no longer happens? | The end state is filled in | Example end state per vertical, to adapt |
| J7 Define the outcome | Answers the Outcome Canvas questions for each problem they build now: user, decision, what slows it down, success, measure | The canvas, prefilled with the user role and key question from the problem cards | All questions are answered | Example answers for the problem |
| J8 Define value in 30-60-90 | Fills in the three horizons and assigns the chosen problems to them | Per horizon: what is delivered, for whom, the expected value, how it is measured, and whether it uses AI | All three horizons are filled in, and the 30-day step contains the problems they build now | Example 30-60-90 per vertical |
| J9 Prepare the build | Reviews the generated starter prompt and copies it into the AI App Builder | A starter prompt built from J6 to J8, and the build folder name to use for this app | The prompt is in the AI App Builder | Use the starter prompt as is |
| J10 Build the app | Builds and refines the app with the AI App Builder | Checkpoints per problem: the app shows data, answers the key question, and supports the decision from J7 | The participant confirms the app answers the key question of every problem they build now | Ask a trainer, or skip to J13 and choose the reference app |
| J11 Upload the app | Uploads the zipped build output of their app | Validation, publication to the DataMiner public folder, and a link to open the app | The app opens through the link | Re-upload after fixing the build folder name |
| J12 Feedback on session 3 | Rates the session and answers the feedback questions | Short questions, see Feedback Questions | The feedback is saved | – |
| J13 Choose how to continue | Chooses Continue with my app or Start from the reference app | The two paths, explained (see Two Ways into Session 4). The reference app becomes available when the trainer reaches this step. | A path is chosen | Start from the reference app |
Important
J9 gives the participant the build folder name before they build. The app must be built for that folder, otherwise it does not load after upload. See App Upload.
Session 4 – Amplify
Follows the Specialized Agents Workshop session flow. The companion carries over the problem, end state, and app from session 3.
| Step | Participant does | Companion provides | Done when | Catch-up |
|---|---|---|---|---|
| J14 Start from your app | Opens their own app or the reference app | The link to the latest app version; for the reference app, also its source to open in the AI App Builder | The app opens | Start from the reference app |
| J15 Try the DataMiner Assistant | Asks the Assistant questions about their backend | Suggested questions for their problem, from simple to advanced | At least three questions asked | – |
| J16 Find the knowledge | Selects the organizational rules that apply to their problem | The knowledge pack of the vertical, filtered on the chosen problem | At least one rule is selected | The rule used by the reference agent |
| J17 Design the agent | Fills in the agent design card, including the integration technique | The design card: role, problem, knowledge used, overview content, integration technique, guardrails, test situations; and the AI integration techniques | The design card is filled in | Example design card for the problem |
| J18 Build and test the agent | Builds the agent and checks the overview against the live situation | A test checklist based on the scenario, e.g., an alarm that crosses an escalation threshold during the session | The overview applies at least one rule and adapts to the live situation | Open the reference agent |
| J19 Stretch: integrate and upload | Integrates the overview into the app and uploads a new version | Publication of the new version under the same link; previous versions are archived | The new version opens through the link | – |
| J20 Feedback on session 4 | Rates the session and answers the feedback questions | Short questions, see Feedback Questions | The feedback is saved | – |
End – Summary and Submit
Happens at the end of session 4. In the online format, participants submit whenever they are ready. In session 5, the trainer reminds participants who haven't submitted yet.
| Step | Participant does | Companion provides | Done when | Catch-up |
|---|---|---|---|---|
| J21 Review the journey | Reviews the summary, names the next increment, and gives overall feedback | The full journey summary | The next increment is filled in | – |
| J22 Submit | Confirms consent and submits | Sends the journey summary document to the training owner (jarne.vynckier@skyline.be), and a copy to the participant if they entered an email address |
The companion confirms that the summary was sent | Download the summary document and send it manually |
After submitting, the companion stays available. Participants can keep working towards their end state and submit an update later.
Feedback Questions
The questions measure whether the training achieved its success measures. They replace a separate feedback form.
After session 3 (J12):
- How useful was this session for your daily work? (1–5)
- Can you name the decision your app supports? (yes/partly/no)
- What did you find hardest: defining the outcome, understanding the data, or building the app?
- What should we change?
After session 4 (J20):
- How useful was this session for your daily work? (1–5)
- Do you see why an agent needs organizational knowledge, not only data? (yes/partly/no)
- Which integration technique fits your organization today?
- What should we change?
Overall (J21):
- Can you explain the four key motions and 30-60-90 to a colleague? (yes/partly/no)
- How likely are you to continue towards your end state in the next 30 days? (1–5)
- What would help you continue?
- Any other feedback?
Journey Summary
The summary is shown in J21 and sent in J22. It contains:
- Participant: display name, company, email (if entered), event, and date
- Vertical and problems: the chosen problems, with the ones built now marked
- End state: as defined in J6
- Outcome: the Outcome Canvas answers from J7
- Value in 30-60-90: the three horizons from J8, with the problems assigned to them
- Apps: the link to the latest version, every uploaded version with its upload time, and whether it started from the reference app
- Agent: the agent design card from J17, the integration technique, and whether the overview was integrated
- Next increment: what the participant will do next, and why
- Feedback: all answers from J12, J20, and J21
- Progress: which steps were completed and which catch-up shortcuts were used
Note
App links point to the participant's DataMiner System. The training owner can only open them with access to that system. How to view apps after the event is an open decision.
Journey Summary Document
The email contains the summary as a Markdown document (empower-journey-<event>-<display name>.md) with a fixed structure. Every summary has the same front matter and headings, so the training owner can give one or many summaries to an agent to review, compare, and aggregate them, for example, "Which problems were chosen most, and which value did participants expect at 30 days?"
Rules for the document:
- The front matter contains all values that can be counted or filtered.
- Headings and their order never change. Empty answers are written as
_Not answered_, never left out. - Free text from the participant is kept as written, under its own heading.
- Ratings use the same scale everywhere (1–5).
Template:
---
type: empower-journey-summary
version: 1
event: Stockholm 2026
date: 2026-11-18
participant: Jane Doe
company: Example Telco
vertical: Telco
problems_chosen: [Priority Dashboard, Customer Impact View, Capacity Risk Workspace]
problems_built: [Priority Dashboard]
level: starter # starter (1 problem built) · intermediate (2) · advanced (3)
session4_path: own-app # own-app · reference-app
app_versions: 2
agent_built: true
agent_integrated: false
integration_technique: Suggest button
steps_completed: 21
catch_up_used: [J6 example end state]
rating_session3: 4
rating_session4: 5
likelihood_to_continue: 4
submitted_at: 2026-11-18T16:42:00
---
# Empower Journey – Jane Doe – Telco
## End State
...
## Outcome – Priority Dashboard
<!-- one Outcome section per problem built -->
| Question | Answer |
|---|---|
| Who is the user? | ... |
| Which decision do they need to make? | ... |
| What slows that decision down today? | ... |
| What does success look like? | ... |
| How would we measure the value? | ... |
## Value in 30-60-90
| Horizon | Delivered | Problems | Expected value | Measured by | Uses AI |
|---|---|---|---|---|---|
| 30 days | ... | ... | ... | ... | No |
| 60 days | ... | ... | ... | ... | Yes |
| 90 days | ... | ... | ... | ... | Yes |
## App
- Latest version: <link>
- Versions: ...
- Started from: own app
## Agent
- Role: ...
- Knowledge used: ...
- Overview content: ...
- Integration technique: ...
- Guardrails: ...
## Next Increment
...
## Feedback
### Session 3
...
### Session 4
...
### Overall
...
Tip
To review an event, collect the summary documents in one folder and ask an agent to summarize them by vertical, level, and session 4 path, and to list the most common feedback. The fixed front matter makes this reliable.