AI is now standard in Swiss social media teams
AI isn't a side project anymore. A new study shows that nearly all large Swiss companies and organisations have plugged AI tools into their social media workflows to boost reach, visibility, and-ironically-trust.
The research, run by ZHAW's Institute of Applied Media Studies with Bernet Relations, examined the 96 largest organisations in the country and contacted 435 overall in autumn 2025.
What the study found
- AI is used mostly for writing, editing, summarising, and translations.
- Adoption is lower for image generation and topic discovery.
- Traditional content management still dominates day to day work.
- Monitoring and evaluation take priority, while dialogue and relationship-building get less attention.
Or as Guido Keel, head of the Institute of Applied Media Studies at ZHAW, put it: "Today, the majority of networks are purely entertainment and consumer platforms."
Why this matters for PR and communications
AI is speeding up content ops, but it's also nudging teams toward volume over conversation. That's a trade-off. If dialogue and contact management fade, community trust will follow.
The upside: faster multilingual output, consistent tone, and better coverage. The downside: sameness, weaker authenticity signals, and less direct contact with the people who actually move your reputation.
Top concerns you need to manage
- Authenticity: Audiences doubt what's real. Overuse of templated AI copy or synthetic visuals makes it worse.
- Public criticism: AI misfires spread quickly, especially when context is off or culturally sensitive topics are simplified.
- Data risk: Copying confidential material into third-party tools invites trouble. So do unclear retention policies.
What to do next (practical checklist)
- Codify use cases: Approve AI for drafts, summaries, translations, and repurposing. Require human review for anything published.
- Label smartly: If AI meaningfully shaped a post (or image), disclose it in a simple, consistent way. It builds credibility.
- Build a prompt library: Store tested prompts for brand voice, tone per channel, crisis responses, and localization. Update monthly.
- Guardrails for images: Limit synthetic visuals to clearly appropriate contexts; avoid human likenesses unless legally cleared.
- Topic discovery: Pair AI with human editors. Use AI to surface angles, then validate with audience insights and media trends.
- Language QA: Run translations through AI + native reviewer. Publish only after a human signs off.
- Security policy: Block uploads of sensitive data to public tools. Use enterprise accounts, log prompts, and set retention limits.
- Rebalance the "social": Set weekly targets for replies, DMs, and community features. Assign owners, not "everyone."
- Crisis drills: Simulate an AI error and rehearse takedown, correction, and response within 30-60 minutes.
What to measure
- Trust signals: Ratio of positive/neutral comments, repeat engagers, and share-to-like rates.
- Dialogue health: Response time, response rate, and percentage of posts with genuine back-and-forth.
- AI impact: Time saved per asset, translation turnaround, content error rate before vs. after AI review.
- Risk events: Number of AI-related corrections, takedowns, or privacy incidents-and time to resolve.
Strategic takeaway
AI is now table stakes for social media production. The advantage goes to teams that pair speed with clear editorial judgment and real conversation. Tighten your guardrails, keep humans in the loop, and put community interaction back on the agenda.
If your team is upskilling on AI for social content and campaign workflows, see our AI certification for marketing specialists for structured, hands-on practice.
Your membership also unlocks:
