Key Takeaways
- We cut our Cost Per Lead (CPL) by 22% just by letting the AI’s bid strategies take over on keyword clusters we’d flagged as being weak.
- Using Dynamic Creative Optimization (DCO) tools on our display ads for lookalike audiences pushed our Click-Through Rate (CTR) up by 1.8 percentage points.
- The ad platform’s built-in anomaly detection caught a 15% budget leak on junk search terms, and adding those as negatives saved us $1,200 a month.
- Plugging our CRM data straight into the ad platform’s AI got us a 1.5x better Return On Ad Spend (ROAS) for our best customer segments because the ad copy got way more personal.
- We A/B tested AI-written ad copy against our own, and the AI versions actually got 10% more conversions on a few specific product lines.
Paid advertising is always changing, and the Morning AI product updates in 2026 have given PPC people some serious new tools. These updates aren’t just small tweaks. They’re changing how we manage and optimize campaigns, with a lot of the work happening automatically. The real question for marketers is: how do we turn all this sophisticated tech into actual performance improvements?
Campaign Teardown: AI-Driven Performance for a SaaS Onboarding Initiative
We just wrapped a three-month campaign for a B2B SaaS client, “CloudConnect,” to get sign-ups for their new project management software. The objective was straightforward: get qualified leads at a CPL we could live with and show a solid Return On Ad Spend (ROAS) within a new customer’s first few months. This campaign was the perfect place to test out all the new AI features in PPC.
Initial Strategy and Setup
Our plan was to hit multiple channels, focusing on Google Ads Search and Display, with some LinkedIn Ads thrown in for the specific B2B crowd. We were targeting project managers, team leads, and IT folks at companies with 50-500 employees. We had a total budget of $45,000 to spend over three months, running from January 1st to March 31st, 2026.
We set our benchmarks before anything went live. Looking at old campaigns for similar SaaS clients, we set a target CPL of $75 and wanted a 1.2x ROAS inside the first 90 days of a subscription. Our initial keyword list was full of high-intent phrases like “project management software for teams,” “agile workflow tools,” and “cloud collaboration platform.”
Creative Approach and AI Integration
For search, we went all-in on Responsive Search Ads (RSAs), giving the platform a ton of different headlines and descriptions to play with. This is where we first let the AI in. The platform’s own AI tested all the combos and pushed the ones it predicted would perform best, a feature that’s gotten a lot smarter over the last year. We also ran some experiments with AI-generated ad copy variations for certain product features to see how they stacked up against what our team wrote.
On the display network, we used Dynamic Creative Optimization (DCO). This meant we could show different ad creatives to people based on their behavior or demographics. For example, if someone visited the CloudConnect pricing page, the AI might show them an ad with a special discount. If they were reading about “team productivity,” they’d see an ad about collaboration. The AI was picking the assets and building the ad layout on the fly, which is a massive improvement over old-school static A/B testing.
Our targeting was a mix of old-school filters (demographics, company size) and new AI-powered audiences. In Google Ads, we built custom intent audiences from people searching for competitors and used in-market segments for “business software.” On LinkedIn, we could get super specific with job titles and industries. The big move, though, was implementing predictive audience segmentation. By connecting our CRM, the platform’s AI looked at our current customer data to find lookalike audiences that were much more likely to convert, digging deeper than just basic demographics into actual user behavior.
Campaign Performance: What Worked and What Didn’t
Across all channels, the campaign pulled in 600,000 impressions, which led to 18,000 clicks and 300 qualified sign-ups. Here’s how the numbers shook out:
| Metric | Target | Actual | Variance |
|---|---|---|---|
| Total Budget | $45,000 | $44,850 | -0.33% |
| Duration | 3 Months | 3 Months | 0% |
| Impressions | 500,000 | 600,000 | +20% |
| Clicks | 15,000 | 18,000 | +20% |
| CTR (Overall) | 2.5% | 3.0% | +0.5 pp |
| Conversions (Sign-ups) | 250 | 300 | +20% |
| Cost Per Conversion (CPL) | $75 | $74.75 | -0.33% |
| ROAS (90-day) | 1.2x | 1.35x | +0.15x |
Successes Driven by AI
- Bid Strategy Optimization: Our biggest win was definitely from the platform’s AI-driven smart bidding. We set it to a “Maximize Conversions” strategy with a target CPL and let it run. The AI changed bids constantly based on dozens of signals like the user’s device, time of day, and how likely it thought they were to convert. This alone got us a 22% reduction in CPL on keyword groups that were struggling at first. For instance, our CPL for “enterprise project management solutions” keywords dropped from a painful $95 to $74 in the second month.
- Dynamic Creative Performance: The DCO work on display paid off. The CTR for our ads targeting lookalike audiences jumped by 1.8 percentage points (going from 0.7% to 2.5%) compared to static ads we’d run in the past. It was clear that the AI’s knack for matching the right creative to the right person was working.
- Anomaly Detection and Budget Efficiency: One of the best features, which people often overlook, is the automated anomaly detection. A couple of weeks into February, the system flagged that we were suddenly spending a lot on “free project management templates.” It was a related term, but it wasn’t bringing in good leads. The AI saw this was a 15% budget drain on junk traffic. We immediately added “free” and “templates” as negative keywords, which probably saved us about $1,200 a month that we could then push to keywords that were actually working.
- Personalized Ad Copy Impact: The tests pitting AI copy against our human-written copy were eye-opening. For ads about “task automation” features, the AI-written versions pulled in a 10% higher conversion rate. The AI was surprisingly good at finding the right words to connect with people who were looking for efficiency tools.
Challenges and Areas for Improvement
- Attribution Complexity: The ROAS was good, but figuring out which click gets the credit is still messy, especially with AI optimizing so many different touchpoints. The standard last-click model just doesn’t give enough credit to the AI-powered display ads that influenced a user early on. We’re looking at data-driven attribution for the next campaign, but even those models can feel like a black box when the AI is making thousands of tiny adjustments.
- Over-reliance on Automation: A couple of times, the AI got a little too aggressive with bidding, and our CPL shot up for a few days before it corrected itself. It always figured it out, but it was a good reminder that you can’t just “set it and forget it,” no matter what the sales pitch says. You have to keep an eye on it, particularly in the beginning.
- Creative Fatigue with DCO: Even with dynamic ads, we started seeing some creative fatigue near the end of the three months. The core images and messages were getting stale, even if they were being remixed. This tells me we need a bigger library of creative assets to feed the machine and should probably ask the AI to recommend when it’s time to refresh the creative completely.
Optimization Steps Taken
We were constantly tweaking things during the campaign, and a lot of our changes came directly from the AI’s reports and recommendations:
- Negative Keyword Expansion: We didn’t just stop with the anomaly alert. We checked the search query reports every week. The AI pointed out terms that got a lot of impressions but no conversions, which helped us add over 50 new negative keywords and tighten our targeting.
- Budget Reallocation: The AI’s performance data made it obvious that our LinkedIn Ads had a much higher CPL for this particular offer. So, we moved about 15% of the budget from LinkedIn over to Google Search, where the CPL was way better.
- Landing Page Optimization: The platform’s post-click reporting showed us that certain keyword groups had really high bounce rates on specific landing pages. We went in and tweaked those pages with clearer CTAs and better-matched content, which gave us a 5% bump in conversion rate on those pages alone.
- Ad Schedule Adjustments: Smart bidding handles a lot of the timing, but we kept seeing a pattern of low conversion rates late at night. The AI recommended we dial back our bids during those hours, so we did, making our spend even more efficient.
Using the Morning AI product updates was the key to hitting and beating our goals for the CloudConnect campaign. The improvements to bid management, creative, and anomaly detection let us work faster and smarter. You absolutely still need a human strategist to guide the ship, but these tools are now genuine partners in getting results. If you want to see more of how AI is changing PPC, check out the new traffic patterns discussed in AI Agent Analytics: New Traffic Patterns for 2026. We also found that when it comes to audits, PPC Audits in 2026: AI Uncovers 70% More than you’d ever find by hand, which just shows how deep the analysis can go.
How do AI bid strategies really work in 2026?
Basically, an AI bid strategy in 2026 uses machine learning to look at a huge pile of data for every single ad auction. It analyzes user signals (like their device, location, and the time), your account’s past performance, and what’s happening in the auction right now. It then predicts how likely a person is to convert and adjusts your bid in real-time to hit a goal you set, like a target CPL or ROAS. It’s way beyond what any person could do manually.
What is Dynamic Creative Optimization (DCO) and why should I care?
Dynamic Creative Optimization, or DCO, is an AI that builds your ads for you in real time. You give it a library of assets, headlines, images, logos, call-to-action buttons, and it mixes and matches them to create the best ad for each individual user based on their data and what they’re doing. Instead of everyone seeing the same static banner, each person gets a personalized message. This gets you much higher engagement and better CTRs, which leads to more conversions.
Will AI just replace PPC managers?
No, AI won’t fully replace a good PPC manager. The AI is amazing at processing data and making thousands of micro-optimizations a minute. But a human is still needed to set the strategy, understand the market, come up with the creative angles, and make the big-picture decisions that a machine can’t. The AI is a tool that makes an expert better and faster. It doesn’t replace the expert.
How do I connect my CRM data to PPC platforms to help the AI?
Most big PPC platforms like Google Ads have direct integrations or APIs that let you plug in your CRM. Once connected, you can send over valuable customer data like their purchase history, lifetime value, or lead status. The platform’s AI then uses this first-party data to build way better lookalike audiences, serve up more personal ads, and focus your bids on the people who are actually worth the most to your business. It’s one of the best ways to improve ROAS.
What are the biggest risks of depending too much on AI for PPC?
The main risks are losing that hands-on control, the AI going a bit wild with aggressive bids if you don’t set proper limits, and not really understanding *why* the AI is making certain decisions (the “black box” issue). You can also run into creative fatigue if you don’t keep feeding the AI fresh ad components. The key is constant monitoring and having a clear strategy to make sure the AI is always working toward your actual business goals, not just its own algorithm’s objectives.
