Artefact Prova
AI Builder Reality Check pour marketers
Décide si tu es prêt à construire une première tranche utile de produit IA, et identifie ce qui cassera en premier si tu démarres maintenant.
Champs du template
- User and painful job
- First useful slice
- Proof that someone wants it
- Current artifact or prototype
- Weakest assumption
- Data or workflow dependency
- Review standard
- Launch risk
- Next 14-day build commitment
Exemple rempli
Les exemples remplis ci-dessous restent en anglais, car Prova évalue les artifacts soumis en anglais.
A marketer trying to build an internal campaign QA assistant.
Version faible vs version solide
Version faible
| User | Marketing team |
|---|---|
| First useful slice | AI assistant for campaign QA |
| Proof | Everyone says QA is annoying |
| Weakest assumption | Need better prompts |
| Launch risk | Bugs |
| 14-day commitment | Build MVP |
Pourquoi cela échoue
- "Marketing team" is not a user.
- "AI assistant" is too broad.
- The proof is hearsay, not behavior.
- The weakest assumption ignores source data, approval rules, and workflow ownership.
- "Build MVP" is not a commitment.
Version solide
| User | Paid media manager checking campaign launch settings before client approval |
|---|---|
| First useful slice | A checklist reviewer that compares campaign setup notes against required client launch rules |
| Proof | Two managers already use a manual spreadsheet before every launch; mistakes still reach strategist review |
| Current artifact | One sample launch checklist, three anonymized campaign setup notes, and one failed QA example |
| Weakest assumption | The checklist rules are explicit enough for AI to evaluate without platform access |
| Data/workflow dependency | Client launch rules, platform screenshots, naming conventions, approval owner |
| Review standard | Flag missing budget, geo, naming, tracking, claim, and approval evidence; never approve launch automatically |
| Launch risk | False confidence before a client-visible campaign goes live |
| 14-day commitment | Build a reviewer for one client and one platform, then test against five past launch packets |
Pourquoi cela fonctionne
- The user is specific.
- The first slice is constrained.
- Proof comes from an existing workflow.
- Launch risk is clear.
- The commitment creates evidence, not just more code.
Ce que Prova vérifie et que l’IA générique rate souvent
- Whether the first useful slice is still too large
- Whether the user is real or imagined
- Whether there is proof beyond personal excitement
- Whether the artifact standard exists before the build starts
- Whether the hard part is product judgment, data access, approval, or implementation
- Whether the next step should be reality check, build brief, build plan, execution lane, or launch gate
Prochaine étape
Tu veux un retour sur ta version? Prova commence par une courte évaluation pour aligner le standard de revue avec ton rôle, ton objectif et ta première audience. Ensuite, tu entres dans le sprint adapté à ton travail actuel.
Prova est actuellement disponible uniquement en anglais.
Avant d’envoyer: retire les noms de clients, les chiffres confidentiels et tout élément que ton équipe ne voudrait pas stocker dans un système de formation ou de coaching.
Continuer vers Prova en anglais