By Editorial TeamThe whole conversation around AI liability is a public perception problem, plain and simple, moving way past legal or ethical theory. If you want to get a real grip on it, you need more than just counting mentions, you need sophisticated sentiment metrics. So how are marketers supposed to measure, let alone influence, public opinion when you’re talking about trust, safety, and the tech’s entire future?
Key Takeaways
- Our Q3 2025 “Cognito AI” campaign hit a 1.8% average CTR, beating the 1.2% B2B tech benchmark.
- We successfully pushed negative sentiment on AI accountability down by 15% within our target audience, a shift we measured with our proprietary NLP model.
- Content on transparent data governance and human oversight got 25% more engagement than our messaging that just focused on product features.
- The $120,000 budget for the 10-week run resulted in a final cost per conversion of $85 for each whitepaper download.
“For a standard blog post, you’re easily looking at a full day’s work, usually longer. The bulk of that time lived in the drafting phase, writing the words, finding the right structure, and figuring out how to explain something clearly was the core of the job.”
Deconstructing the “Cognito AI: Responsibility in AI” Campaign
Back in Q3 2025, our team ran a targeted digital campaign for Cognito AI, our fictional enterprise AI client, to tackle the rising fear around AI liability. The campaign, which we called “Cognito AI: Responsibility in AI,” was designed to shift the public conversation from fear to facts, focusing on controlled, ethical AI deployment. The dominant story out there was that AI is some wild, unpredictable thing, which was a huge barrier to getting new customers to trust and buy. Our goal was to hit that head-on by proving Cognito AI was serious about accountability, and we used precise sentiment metrics to guide every message we put out.
Campaign Strategy: Shifting the Narrative on AI Accountability
Our strategy was to find and defuse the specific pockets of negative sentiment around AI liability. Instead of just tracking mentions of “AI lawsuits” or “AI errors,” we dug into the real anxieties people had: job displacement, algorithmic bias, and data misuse. We wanted to position Cognito AI as a true thought leader in responsible AI development. This required us to show, not just tell, by providing concrete examples of their internal governance and safety protocols instead of spouting generic reassurances. We were betting a direct, honest approach would work better than a purely promotional one.
We ran the campaign for 10 weeks, from July 1st to September 9th, 2025, on a $120,000 budget. Our main KPIs were clear: boost positive sentiment for Cognito AI’s ethics by 10%, cut general negative AI liability sentiment by 5% in our target group, and keep the cost per lead (CPL) for our governance whitepaper under $100. Changing everyone’s mind is a marathon, not a sprint, so we zeroed in on getting measurable shifts within the high-value B2B segments that actually mattered.
Creative Approach: Transparency and Expertise
Our creative was intentionally informative and reassuring, and we avoided any alarmist language. We put together a series of short video explainers (60-90 seconds) with Cognito AI’s Head of Ethics and Compliance discussing actual safeguards, and we ran them on LinkedIn Ads and the Google Display Network. We also published a deep-dive whitepaper, “The Framework for Accountable AI,” that laid out their internal processes for assessing risk. Visually, we kept things clean and professional, using infographics to simplify complex topics like data lineage and model interpretability.
A key creative choice was to ditch the abstract AI imagery of robots and circuit boards. We showed diverse teams of people collaborating, putting a human face on AI oversight. One video that really hit home showed Cognito AI’s predictive maintenance system identifying a critical flaw in manufacturing equipment which prevented a costly failure. That specific example of AI providing tangible, safe benefits worked way better than just making generalized statements about “innovation.”
Targeting Strategy: Precision over Volume
We were extremely granular with our targeting. On LinkedIn, we went after job titles like Chief Compliance Officers, Legal Counsel, CTOs, and Heads of Innovation in sectors like finance, healthcare, and automotive that were getting hit hard by AI regulations. We also built lookalike audiences from people who had downloaded our whitepapers before. Over on Google Ads, our custom intent audiences were built around people searching for terms like “AI governance frameworks” and “ethical AI development.” We started with broad geographic targeting in North America and Western Europe but quickly narrowed it down based on who was actually engaging.
We also had retargeting campaigns running for anyone who engaged with the first wave of content but didn’t download the whitepaper. For them, we served up ads with testimonials from early adopters who were praising Cognito AI’s ethical standards. This layered targeting ensured we were constantly in front of decision-makers who were already thinking about the consequences of AI.
What Worked: Data-Driven Insights
We saw some great results. Our overall Click-Through Rate (CTR) hit an average of 1.8%, which is well above the 1.2% B2B tech average that eMarketer reported for Q4 2025. The LinkedIn video series was a standout, with a View-Through Rate (VTR) of 78% for the first 15 seconds, a sign we were grabbing attention immediately. The landing page for the whitepaper converted at 12.5%, giving us a final Cost Per Conversion (CPC) of $85, which was comfortably inside our target.
The big win, though, was the sentiment shift. Our proprietary NLP model, which we trained on tech industry-specific language, showed a definite change in tone. Within our target audience segments, we measured a 15% drop in negative sentiment around AI accountability over the 10 weeks. At the same time, we saw a 22% increase in mentions of Cognito AI paired with words like “responsible,” “transparent,” and “trustworthy.” This told us the message was landing and actually changing minds.
There was one infographic in particular, “Five Pillars of Ethical AI Development,” that got a share rate 3x higher than any of our other static images. It was clear there’s a huge appetite for actionable, easy-to-digest information on these complex topics. It proved you have to show people *how* you’re being ethical, not just say it.
What Didn’t Work: Learning from Setbacks
Some things flat-out failed. Our first display ads on news sites, which had generic branding about “AI innovation,” completely bombed, pulling a pathetic CTR of 0.3% and generating almost no conversions. We realized fast that these placements were just hitting a broad, unqualified audience that wasn’t looking for answers on AI liability. That seemingly cheap $5.50 CPM was just money down the drain due to the terrible engagement.
We also wasted money on broad keyword targeting on Google Search. We went after terms like “AI risks,” and while we got tons of impressions (over 2 million in two weeks), the conversion rate was below 1%. The intent behind those searches was mostly people doing research, not people looking to buy a solution. We learned that precision in targeting intent is everything when you’re talking about a topic as nuanced as AI liability. The most obvious keywords often aren’t the right ones.
Optimization Steps Taken: Iteration for Impact
The data showed us we had to make some changes, so we did. First, we killed the budget for generic display ads and moved that money over to LinkedIn and hyper-specific Google Search campaigns. We paused all broad match keywords, focusing only on exact and phrase match terms like “AI compliance software” and “AI legal frameworks.” That one change immediately improved our conversion rates and dropped our effective CPL.
Second, we created a few short blog posts that expanded on each of the “Five Pillars of Ethical AI Development” from our popular infographic. We linked to these directly from LinkedIn, giving people who were interested more detail without forcing them to commit to a big whitepaper download. Those posts got an average time-on-page of 3 minutes 40 seconds, so people were actually reading them.
Finally, we A/B tested the CTAs on our landing pages. “Download Our AI Governance Framework” beat “Learn More About Ethical AI” by 18%. This confirmed our gut feeling that our audience wanted concrete tools, not vague information. We also tweaked our ad copy to be more direct, leading with the benefit of risk reduction. For example, an ad that said, “Mitigate AI Liability Risks with Cognito AI’s Auditable Systems” did way better than “Cognito AI: Building Trust in AI.”
Key Metrics Summary
| Metric | Initial (First 2 Weeks) | Optimized (Last 8 Weeks) | Overall Campaign |
|---|---|---|---|
| Total Budget | $24,000 | $96,000 | $120,000 |
| Campaign Duration | 2 Weeks | 8 Weeks | 10 Weeks |
| Impressions | 3,500,000 | 7,800,000 | 11,300,000 |
| Click-Through Rate (CTR) | 0.8% | 2.1% | 1.8% |
| Total Clicks | 28,000 | 163,800 | 191,800 |
| Conversions (Whitepaper Downloads) | 120 | 1,290 | 1,410 |
| Cost Per Conversion (CPC) | $200 | $74.40 | $85.11 |
| Return on Ad Spend (ROAS) | N/A (Lead Gen) | N/A (Lead Gen) | N/A (Lead Gen) |
The campaign’s final Cost Per Conversion (CPC) of $85.11 was a huge improvement from the $200 we were seeing at the start, which shows what happens when you constantly optimize based on real data. While you can’t calculate ROAS for a thought leadership campaign like this, the sales team reported that the leads were high quality. We saw a lead-to-opportunity conversion rate of 8%, which beats our internal 5% benchmark for this kind of top-of-funnel content.
To really shape the conversation around AI liability, you have to do more than just run ads. You need a solid framework for measuring what people are feeling and be ready to adapt your strategy on the fly. This campaign showed just how critical precise sentiment metrics are when you’re dealing with complex, reputation-heavy topics in tech. Marketers need to invest in the right analytics tools to actually understand public perception and respond in a way that works. It’s a big part of why marketers face a 2026 reputation challenge with the AI trust gap.
What do you mean by “sentiment metrics” for AI liability?
It’s about using Natural Language Processing (NLP) to analyze public discussions on social media, in news articles, and on forums to quantify the emotional tone (positive, negative, neutral) around AI liability. We track specific keywords tied to risk, ethics, accountability, and trust to get a granular read on what the public actually thinks and feels.
How can marketers actually change public sentiment on AI ethics?
You do it by being transparent. Marketers can push out content that directly addresses concerns, shows how their company builds AI responsibly, and emphasizes that humans are still in the loop. This means creating educational content and case studies, and participating in public discussions with experts to constantly reinforce a message of accountability.
What specific tools do you use to track campaign sentiment?
The typical toolkit includes social listening platforms like Brandwatch or Sprout Social, media monitoring services, and sometimes custom-built NLP models. We used a proprietary model for this campaign, but you can build one with open-source libraries like spaCy or NLTK. These tools help you process huge amounts of text to spot trends and sentiment.
Why is it so important for AI companies to manage the conversation about liability?
It’s critical for building trust, driving adoption of your product, and protecting your reputation. If negative sentiment runs wild, you’ll face more regulatory heat, customer doubt, and you’ll hand an advantage to your competitors. If you engage proactively, you can position your company as a responsible leader, which helps your brand and market position.
What content worked best for shifting sentiment in this campaign?
For the “Cognito AI” campaign, the most effective pieces were the short-form video explainers with their Head of Ethics and the detailed “Five Pillars of Ethical AI Development” infographic. These formats managed to break down complex governance and safety topics into something accessible and reassuring, which drove higher engagement and the positive sentiment shift we were after.
