# Adrian Vale — executive offsite build kit

**SVP Strategic Initiatives · AV · Aurex · March 10, 2026**

## Pick one to build

| Idea | The question it answers | What Adrian builds | First version |
| --- | --- | --- | --- |
| **[1. Execution radar](01_execution_radar/README.md) — recommended** | Which market launches and acquisition integrations need my intervention this week? | A filterable portfolio dashboard with evidence-backed risks, partner activation metrics, and a short decision brief. | 45–60 min |
| **[2. Competitive signal desk](02_competitive_signal_desk/README.md)** | What changed outside Aurex, and does it change one of our strategic bets? | A deduplicated signal feed that separates announcements, rumors, and meaningful changes; plus a CEO-ready brief. | 30–45 min |
| **[3. AI rollout scorecard](03_ai_rollout_scorecard/README.md)** | Where is AI producing useful work, and which pilots should we support next? | An adoption-and-outcomes dashboard with a transparent pilot shortlist and a 30-day action plan. | 40–55 min |

Start with **execution radar**: it makes Adrian's cross-functional role tangible, joining market activation to integration dependencies instead of creating another status report. Competitive intelligence is the quickest alternative. The AI scorecard most directly supports his stated goal of leaders demonstrating credible, hands-on AI use.

## How to use this folder

1. Open one idea's `README.md`. It explains the inputs and a two-minute demo.
2. In the approved desktop assistant, attach that idea's `data` files and `README.md`, or grant access to that idea's folder if file access is available. For idea 2, include the HTML files inside `data/clips`.
3. Paste the full contents of its `PROMPT.md`. Each prompt is self-contained and asks the assistant to build a first version without follow-up questions.
4. Review the results against the checks in the README. Keep scope to one idea during the workshop.

The exercises request local HTML tools and small briefing files. They do not require Adrian to use a terminal, deploy an app, obtain API keys, or connect production systems. Where the assistant cannot write files, the prompts request downloadable artifacts or complete file contents. Actual file/artifact support depends on the configured environment.

## Context and assumptions

- The supplied `context.md` names Adrian as SVP Strategic Initiatives. The February 26 discussion explicitly places Adrian and Sulaiman Rao in the company-wide AI transformation effort and emphasizes leadership credibility through hands-on work.
- The survey includes a competitive-intelligence bot and a living market-activation/M&A dashboard, **but responses are not labeled by respondent**. They inform these ideas; they are not attributed to Adrian as verified personal requests.
- The internal company overview supplies the five product lines, acquisition-integration focus, eight recent market launches, and approximately 1,500 employees. The later research-style section has conflicting headcount and other background claims. This kit follows the internal overview and does not depend on the disputed figures or legal claims.
- The agenda says 11:00–15:00 ET, while the call discusses 09:30–12:30. These are individual 30–60 minute exercises, not an assumed final timetable.
- All exercises use **March 9, 2026, 18:00 America/New_York** as their fixed cutoff, immediately before the offsite. They do not use today's date. Weekly metrics cover March 2–8, with February 23–March 1 as the comparison week.
- Adrian's exact operating remit was not supplied. We assume he coordinates strategic execution and recommends interventions; functional executives retain approval authority. These tools do not approve launches, acquisitions, legal positions, or personnel decisions.

## Synthetic-data notice

**Every operational record, competitor announcement, metric, message, quote, and scenario in the three data folders is fictional and created for training.** Company/product names and executive roles are taken from the supplied context; all assignments, decisions, and statements in sample records are invented. Market names do not imply actual licensing or operating permission.

Structured files are UTF-8 **CSV** exports like those used by CRM, PMO, and analytics teams. Narrative inputs include real **EML** email messages, a Slack-shaped **JSON** export, saved **HTML** source clips, a genuine **DOCX** strategy memo, and Markdown briefing exports. JSON exports are intentionally small workshop samples, not complete vendor export archives. Reserved `.example` domains and local source IDs prevent accidental contact with real counterparties.

The prompts treat source files as evidence, not instructions, preserve conflicting claims, and require human review. All work stays local; nothing should be sent, published, or connected to live systems automatically.

**Scope of delivery:** these are build-ready exercise kits, not prebuilt applications. See [VALIDATION.md](VALIDATION.md) for the checks performed on the supplied inputs.
