AI Emotional Intelligence: 3.5x ROAS in 2025

Back in Q4 2025, we ran a big campaign for a B2B SaaS client, “ConnectFlow.” They have a customer support automation platform, and the goal was to get more demo requests. The real problem was cutting through the noise, everyone’s selling AI efficiency. We decided to go a different route and focus on their unique angle: injecting emotional intelligence in AI customer interactions. The whole campaign hung on one question: could we prove that an AI could actually respond to human sentiment, not just parse keywords?

Campaign Strategy: Beyond Keywords to Connection

Our whole strategy for ConnectFlow was to *show*, not just tell, how its AI could improve customer experience (CX) with genuine emotional understanding. Our bet was that businesses were sick of talking to robots and were ready for an AI that could tell if a customer was frustrated, happy, or in a hurry and actually change its response. This meant going deeper than keyword matching to actually understand the underlying sentiment and adjust the chat’s tone and the solutions it offered.

We ran the campaign for eight weeks, from October 1st to November 26th, 2025, spending a total of $75,000 across multiple channels to really test our hypothesis about emotionally intelligent AI, targeting mid-market and enterprise customer service managers, CX directors, and the IT folks who make the final call.

Creative Approach: Empathy in Every Touchpoint

Our creative was all about showing understanding and responsiveness through our visuals and messaging. We deliberately dodged the typical “tech-bro” imagery of shiny dashboards and instead showed human-centric scenes, like a customer actually smiling after a chatbot interaction or a support agent getting an AI alert about a customer’s sentiment. Our message was simple: “ConnectFlow AI: Understand More, Resolve Better.”

Ad Copy A/B Testing

We ran a ton of A/B tests on ad copy. One set of ads hit the usual efficiency points: “Reduce Resolution Times by 30%.” The other set sold the emotional payoff: “Your Customers Deserve to Feel Heard.” The empathy-focused ads won every time. On LinkedIn Ads, for instance, that copy got a 1.8% average click-through rate (CTR), while the efficiency copy only managed 1.1%. That 0.7 percentage point lift might look small, but it drove a big jump in qualified leads right at the top of the funnel.

Video Content: Showing, Not Just Telling

We also produced a 90-second explainer video that showed ConnectFlow’s sentiment analysis in a real-world scenario. The video simulated a customer interaction where the AI picked up on rising frustration and proactively offered to transfer them to a human agent, even providing a summary of the emotional context for the agent. We pushed it out on LinkedIn and in our email sequences. According to HubSpot’s 2025 State of Marketing report, video drives high engagement, and our 45% view-through rate on LinkedIn certainly backs that up.

Targeting: Pinpointing the Pain Points

We targeted companies and people who were obviously looking for a fix for high customer churn, bad reviews, and agent burnout. Our mix included:

• LinkedIn Audience Targeting: We went after job titles like “Head of Customer Success,” “VP of Customer Experience,” and “Director of Support Operations” at companies with 500+ employees, and also hit members of CX-focused professional groups.
• Google Ads Search Campaigns: We bid aggressively on high-intent, long-tail keywords like “AI sentiment analysis customer service,” “emotionally intelligent chatbot,” “CX automation with empathy,” and “customer frustration detection software.”
• Custom Audiences: We uploaded lists of people who had attended industry webinars or subscribed to CX newsletters.

One of our best moves was building a custom audience from companies that had recently posted job listings for roles like “Customer Experience Specialist” or “AI Solutions Architect.” That told us they were already spending money to improve CX and would be receptive to our message.

What Worked: Data-Driven Successes

The campaign’s success was really driven by our focus on emotional intelligence.

LinkedIn InMail Sequences: Personalized Empathy at Scale

We put 30% of our budget ($22,500) into highly personalized LinkedIn InMail sequences. We didn’t just blast a generic message. Every InMail was tailored using public info about the person’s company, like recent news about their customer service or industry trends we could connect to. Subject lines like, “Are your customers truly feeling heard?” got a 28% open rate and a 7% reply rate, way better than our 15% open/3% reply benchmark for standard InMails. The cost per lead (CPL) was $110, and while that’s not cheap, the leads were fantastic and already bought into the idea of emotionally intelligent AI.

Sentiment-Driven Landing Pages

Our landing pages weren’t static. We included a short quiz, “How emotionally intelligent is your current CX?”, that dynamically changed the page content to highlight the ConnectFlow features that solved their specific problems. That simple bit of personalization paid off, boosting our conversion rate to 12% for visitors from the empathy ads versus only 8% from the efficiency ads, a 4 percentage point difference that directly fueled demo requests.

Google Ads Performance: High-Intent Conversions

Our Google Ads search campaigns were killers for specific long-tail keywords. The term “AI customer sentiment analysis platform” got us an average position of 1.5, a 6.2% CTR, and a CPL of $85. Our $20,000 Google Ads spend pulled in 235 qualified leads and 35 direct demo requests, which worked out to a $571 cost per conversion (demo request). We were happy with that number, since ConnectFlow’s enterprise deals have a high lifetime value.

What Didn’t Work: Learning Opportunities

But not everything worked right out of the gate. Here are the duds and what we learned.

Broad Display Network Targeting

Our first attempt at using the Google Display Network (GDN) with broad targeting based on “business software interests” was a waste of money. We got a ton of impressions (over 1.5 million), but the CTR was a pathetic 0.15% and the CPL was a completely unsustainable $350. The ads looked good, but without the search intent to give them context, they just fell flat. It was a clear lesson for a niche B2B tool like this: context and intent beat broad reach every time.

Generic Email Nurture Flows

Our first email nurture sequence for whitepaper downloads was a miss. It was a generic feature dump instead of continuing the emotional benefit story from the ads. The numbers were bad: 15% open rate, 1% click-through, and a 2.5% unsubscribe rate. We realized that if we hook someone with an empathy message, we have to keep that conversation going through the entire funnel.

Optimization Steps Taken: Iteration for Impact

Based on that early data, we moved quickly to make some key optimizations:

• GDN Refinement: We killed the broad GDN campaigns and shifted 70% of that budget ($7,000) into retargeting people who had watched our LinkedIn video or hit a landing page but didn’t convert. The new retargeting campaign hit a 0.8% CTR and a much more manageable $150 CPL. We also built custom intent audiences on GDN, targeting users who had recently searched for competitor names.
• Email Nurture Overhaul: We rebuilt the entire email flow to feature case studies showing how ConnectFlow directly improved CSAT and NPS scores by understanding emotional signals. We also personalized the CTAs based on what ad they first clicked. For example, if they came from an ad about reducing customer frustration, the emails hammered that point. The new sequence jumped to a 25% open rate and a 3.5% click-through rate.
• AI Model Feedback Loop: On the backend, we worked with ConnectFlow’s product team to set up a real-time feedback loop. Any chatbot interaction where the AI flagged negative customer sentiment got reviewed by our team. This let us see if our marketing was setting the wrong expectations or if we were missing key pain points in our content. This loop let us sharpen our marketing copy while also giving the ConnectFlow product team direct insights for their development roadmap.

Campaign Performance Metrics Summary

When the dust settled, the campaign clearly proved the value of marketing an AI’s emotional intelligence. Here’s the final breakdown:

• Total Budget: $75,000
• Duration: 8 weeks (October 1st to November 26th, 2025)
• Total Impressions: 2,800,000
• Overall CTR: 1.2%
• Total Qualified Leads: 680
• Overall CPL: $110.29
• Total Conversions (Demo Requests): 132
• Overall Cost Per Conversion: $568.18
• Revenue Generated (from closed deals within 3 months): $262,500
• Return on Ad Spend (ROAS): 3.5x

A 3.5x ROAS is a strong return in B2B SaaS, especially with a long sales cycle. More importantly, the leads we got from channels that stressed an emotional connection were just better, converting at higher rates all the way down the funnel. In fact, our sales team saw a 20% higher close rate on calls with leads from the “empathy” campaigns compared to the “efficiency” ones.

What this ConnectFlow campaign drove home for us is that B2B SaaS has turned a corner. Efficiency is still table stakes, but an AI’s ability to actually understand and react to human emotion is the new differentiator. Marketers can’t just throw feature lists at people anymore. We have to show how our tech builds real, meaningful customer relationships. That means spending money on creative that shows empathy, finding audiences who are desperate for better interactions, and constantly tuning everything based on sentiment data. For more on this, check out how AI transforms retail CX in 2026 and what it means for AI boosting agility for marketing leaders.