Digital Marketing This Week: AI Is Only as Smart as What You Feed It

Consumers are pushing back on AI ads that feel hollow. Google’s smartest new features only work if you give them real sales data. And Merchant Center just revealed that incomplete product feeds are invisible to AI shoppers.

Here’s what happened between May 25 and 29, and what it means for how you work.

01. 87% of Consumers Say Great Ads Still Need a Human Touch, But Brands Keep Pulling Humans Out

Canva’s State of Marketing and AI 2026 report dropped this week, and the headline number is hard to ignore: 87% of consumers say the best advertising still needs a human touch.

That alone would be worth noting. But the data goes further.

74% say they are more likely to buy from an ad they believe was made entirely by humans. 70% say they can already tell when an ad is AI-generated, because it feels like it’s “missing its soul.” And 65% say AI ads are so obvious, they’re laughable.

The fatigue is not format-specific. It shows up in social posts, emails, product photos, voiceovers, and articles. Consumers have developed a working radar for content that was produced at scale without real creative direction behind it.

Here is where it gets complicated for brands. 99% of marketing leaders say they plan to increase their AI budgets. Consumers are pulling one way; investment is going the other.

The report’s actual argument is not that AI is the problem. It is that AI without human judgment is. 68% of consumers say they are fine with AI in ads when it makes them more helpful or relevant. Younger audiences largely don’t care how an ad was made, as long as real people were visibly involved.

The line consumers are drawing is between AI as a production engine and AI as a replacement for creative intent. Brands on the wrong side of that line are building a trust deficit that efficiency gains won’t recover.

KEY TAKEAWAY Before your next AI-assisted campaign goes out, check whether human judgment is present at every decision point: the message, the audience, the feeling it should create. Speed is not the differentiator anymore. Direction is.

Source: martech.org, canva.com

02. Meta Launched a Reddit-Style App Called Forum, and the Real Reason Is Your Brand’s AI Visibility

On May 22, Meta quietly dropped a new standalone app called Forum. No press release, no announcement. Social media consultant Matt Navarra spotted it first.

The app pulls all your Facebook Groups into one place, organized around discussions and questions rather than a traditional social feed. You sign in with your Facebook account, post with a nickname, and pick up conversations where you left off. It looks and functions a lot like Reddit.

That similarity is not accidental.

Reddit has become one of the most cited sources for AI chatbot answers, specifically because it surfaces real, experience-based responses that have been vetted by communities through upvotes. AI tools reference the most highly upvoted replies as the most likely relevant answers to user queries.

Forum is Meta’s attempt to build that same infrastructure inside its own ecosystem.

The app includes an AI-powered “Ask” tab where users can post questions and receive answers compiled from discussions across multiple groups. The more people ask questions and engage in group discussions, the more vetted, human-sourced content Meta has to power its own AI responses.

For brands, the implication is direct. Facebook Groups are already where niche communities discuss products, compare services, and ask for recommendations. If those conversations start feeding Meta AI answers the same way Reddit feeds other chatbots, being active and authoritative inside relevant groups is not just community management. It is AI visibility.

Meta has tried a standalone Groups app before and shut it down in 2017. Whether Forum sticks depends on whether users actually want another app, or whether Meta finds ways to push activity there. Either way, the intent behind it is clear.

KEY TAKEAWAY Identify the Facebook Groups most relevant to your brand or category and audit how present you actually are in those conversations. Forum is early, but the pattern it is following, community content feeding AI answers, is already proven. The brands that build genuine authority in group discussions now will have a structural advantage when Meta AI starts surfacing those answers at scale.

Source: socialmediatoday.com

03. Google Is Merging Display Network Into Demand Gen: One Place to Run It All

Google announced this week that Display Network ads are moving into Demand Gen, creating a single unified place to manage campaigns across both properties.

Here is what that means in practice.

Google’s Display Network reaches over 90% of global internet users across more than 2 million sites, videos, and apps. Demand Gen covers YouTube, YouTube Shorts, Discover, Gmail, and Google Maps. Until now, these have run as separate campaign types with separate management. Starting June 2026, advertisers can begin migrating Display campaigns into Demand Gen voluntarily. Automatic migration follows later, with the full transition expected to complete by 2027.

The practical upside is real. Google’s own data shows advertisers who added Display inventory to Demand Gen campaigns saw an average 9.5% increase in ROI. One case study, food delivery platform GoFood, saw a 24% drop in CPA and 19% higher conversion volume after making the switch.

The migration is not forced overnight. Eligible advertisers will get a migration tool in their accounts next month. Display-only campaigns can still serve exclusively on the Display Network if that is the preference. The change is about management consolidation, not inventory restriction.

What it does remove is the friction of running two separate campaigns to cover the same customer journey. Discovery intent on YouTube and browsing intent across third-party sites can now be addressed from one campaign structure, with channel controls at the ad group level to manage where each asset type actually serves.

KEY TAKEAWAY When the migration tool appears in your account in June, don’t ignore it. Run a small test combining your existing Display assets inside a Demand Gen campaign and measure the ROI delta before committing fully. The efficiency gains are documented, but your account’s specific audience mix will tell you how far to lean in.

Source: socialmediatoday.com, blog.google

04. Google Just Turned Search Ads Into Conversations, But Only If You Feed It the Right Data

Google’s biggest lead generation announcement this week was not a targeting update or a bidding feature. It was a fundamental change to what a Search Ad actually does.

The new Business Agent for leads lets prospective customers ask questions directly inside a Search Ad, covering pricing, availability, expertise, and services, and receive answers pulled from the business’s own website content. No click to a landing page. No form to fill. The conversation happens inside the ad itself.

Google’s pitch is that the lead who arrives after that exchange is a fundamentally different prospect from someone who clicked a headline and bounced. They’ve already asked questions. They’ve already been qualified. The intent is deeper before they ever enter your pipeline.

That is genuinely valuable, at least in theory.

The problem is the same one that has plagued every AI-driven campaign feature Google has released in recent years. These systems optimize toward whatever signal you give them.

“AI doesn’t have a quality problem. Marketers have an input problem. Feed it form fills, it’ll get you form fills. Feed it closed revenue, and it’ll get you customers. The tool hasn’t changed. What you’re training it on has.”

Gursharan Singh, Co-founder, WebSpero Solutions

Google’s own announcements make this explicit. Nearly every new feature, including lead intent scores, journey-aware bidding, offline conversion imports, and CRM integrations with Mailchimp, Klaviyo, and ActiveCampaign, is built around one premise: the AI is only as smart as the downstream data you feed back into it.

Businesses connecting actual sales outcomes to their campaigns are the ones Google’s systems will increasingly favor. Those still tracking form submissions as their final signal are training the algorithm to optimize for a metric that does not reflect their business.

KEY TAKEAWAY If you run lead generation on Google Ads, your most important task right now is not testing the conversational ad format. It is getting your offline conversion data, specifically sales qualified leads, closed deals, and revenue, flowing back into your campaigns. That is the signal the new AI systems are built to use. Without it, you are paying for a smarter system that is still optimizing for the wrong thing.

Source: searchengineland.com, blog.google

05. Google Is Turning Merchant Center Into an AI Visibility Platform, and Product Feeds Are the New SEO

Google announced this week that Merchant Center is getting a new layer of AI performance insights, giving retailers visibility into how their products appear across AI Mode, AI Overviews, and the Gemini app.

Until now, Merchant Center has told you how products performed in traditional Shopping. What it could not tell you was whether your products were showing up when someone asked Gemini for a recommendation, or how your brand compared to competitors inside an AI Overview result. That gap is what this update addresses.

The new reporting includes four things worth paying attention to.

Share of voice benchmarks your brand’s visibility across AI shopping surfaces against similar retailers, marking the first time Google has offered this kind of competitive context inside an AI environment. Shopping funnel performance breaks down how products are discovered, evaluated, and purchased across conversational surfaces. Product term insights show what conversational queries shoppers are using and where your share of voice stands for each. And product attribute insights identify exactly which structured data fields, including color, material, style, and size, are missing from your feed and likely costing you visibility.

That last one matters more than it sounds. As shopping searches become increasingly conversational, the way Google’s AI surfaces products looks less like a keyword auction and more like an SEO relevance match. A complete, well-structured product feed is becoming the equivalent of a well-optimized page. Incomplete attributes are the equivalent of thin content — and the AI has no way to recommend what it cannot fully understand.

The rollout is limited for now — US, Canada, Australia, India, and New Zealand in the coming months — but the direction is clear. Merchant Center is quietly evolving from a feed management tool into an AI commerce optimization platform.

KEY TAKEAWAY When AI performance insights become available in your account, start with the product attribute completeness score. Missing structured attributes are the fastest fix with the most direct impact on AI discoverability. Treat your product feed the way you treat on-page SEO — the more complete and specific the data, the better the AI can match your products to the right queries.

Source: searchengineland.com, blog.google

 

---

We publish a weekly roundup of the most important updates in SEO, paid search, content strategy, and AI-driven marketing — every week, no noise.