Political Tech: How One Ad Maker Is Thinking About AI and Campaign Creative
Publitics is rebranding to Flashpoint Strategies, refocusing on TV, direct mail, and digital advertising. Leading the politics practice is Henry de Koninck, a veteran creative who's betting on a human-led, AI-assisted model.
The firm's core engine is InfluenceIQ - an AI modeling and performance data platform built to refine creative and place media with precision. Or as de Koninck put it, it's "the tech to ensure the right message reaches the right audience on the right channels."
What InfluenceIQ Means for Creatives
AI isn't replacing talent here - it's sharpening it. The platform feeds performance data back into the creative process so teams can iterate faster, test smarter, and spend where attention actually converts.
De Koninck is clear about the hierarchy: "Is it the magician or the wand? The tech helps us optimize, but it's really the human touch that drives it." Tools scale output; taste sets the bar.
The Line Between Useful and Risky
Campaigns have taken heat for AI use - especially mimicking opponents. Public sentiment isn't exactly optimistic either; polling shows Americans are more concerned than excited about AI in politics. See the data from Pew Research Center.
That's why the operator matters. Without a skilled human, output can be "at best pedestrian and at worst really kind of detrimental." Or, as de Koninck frames it, the creative's taste is the filter: what gets made - and aired - still depends on someone with judgment.
Practical Takeaways for Creatives
- Start with a clear creative thesis. Define your emotional beat and the one-liner the audience should remember before you touch a tool.
- Use AI for options, not answers. Generate variants, then apply your lens to cut 90% and refine the 10% that's actually interesting.
- Treat performance data like a collaborator. Build tight feedback loops (hook, line, CTA, placement) and iterate daily, not monthly.
- Keep a "human veto." If a concept feels off-brand, ethically fuzzy, or likely to backfire, kill it - even if the metrics look tempting.
- Disclose synthetic elements when relevant and avoid impersonation. Follow emerging norms such as Responsible Practices for Synthetic Media.
A Simple Creative + AI Workflow
- Brief: Audience, problem, belief shift, single promise.
- Concept: 5 hooks, 3 angles, 2 narrative frames (human-crafted).
- AI Assist: Variant scripts, alt CTAs, visual references, voice options.
- Human Edit: Tone, ethics, brand guardrails, narrative discipline.
- Pre-Flight: Fact check, legal/brand review, disclosure if needed.
- Launch + Learn: A/B hooks and placements; optimize based on watch time, recall, persuasion lift, and cost per outcome.
Metrics That Actually Help You Improve
- Attention: 3-second hold, 50% view-through, and completion rate.
- Memory: Hook recall and message agreement in quick-turn panels.
- Action: Click-through to priority destinations and cost per desired action.
- Quality: Sentiment and "would you share this?" responses from target segments.
Guardrails Worth Writing Down
- No impersonation of real people; no synthetic quotes. Label alterations.
- Keep an audit trail: prompts, assets, edits, approvals.
- Use licensed or original assets; verify training data rights for any generators.
- Stress-test claims with fact-check and legal sign-off before flight.
AI can crank out volume. But volume without taste is noise. De Koninck's point lands: tools are useful, the creative lens is decisive.
Level Up Your AI Skill Set
If you're building a smarter creative stack, explore curated AI courses by role at Complete AI Training. Learn prompts, workflows, and measurement that keep the human in charge - and the work effective.
Your membership also unlocks:
