How to use this scorecard
Score each of the ten categories from 1 (not in place) to 5 (in place, working, and accountable). Add the totals; the band tells you where you sit and where to invest first.
The ten categories
Trusted knowledge
Does AI have access to the campaign’s real sources of truth — plan, budget, message guidance, donor history — and only those?
Workflow clarity
Are the specific tasks AI is allowed to do named, scoped, and documented, or is it ad-hoc?
Human review
Are the human checkpoints explicit and proportional to the risk of each task?
Decision ownership
For every kind of AI-assisted output, is there a named role accountable for the final call?
Risk controls
Are there clear rules for sensitive data, public-facing content, legal escalation, and incident response?
Staff training
Have staff received practical, role-specific training on the AI workflows they are expected to use?
Adoption
Are the AI workflows actually being used by the people they were designed for, or are they shelfware?
Measurement
Is the campaign measuring whether AI workflows save time, raise money, or improve quality — not just usage?
Tool integration
Do the AI tools connect to the systems staff already work in, or do they create another silo?
Leadership accountability
Is there a senior leader explicitly accountable for AI strategy, governance, and outcomes?
Scoring bands
Related frameworks
- The Campaign AI Operating System — the underlying model the scorecard maps onto.
- Model AI Use Policy for Democratic Campaigns — the operational rules that turn a high score into protection.