B2B AI Buying, Content Gaps & Smart Bidding

Three things shifted at once, and your dashboard missed all of them

B2B buyers now start vendor research with a ChatGPT or Gemini prompt, not a Google search. At the same time, Google is changing Smart Bidding behavior on August 17 in a way that will catch many advertisers off guard. And more than 50% of newly published web content is now AI-generated, meaning the information pool that AI tools draw on when answering your buyers' questions is filling up with synthetic content. Each of these trends deserves its own deep-dive. Together, they change what's worth doing this week.

Key takeaways

• 66% of B2B professionals use AI to research vendors, but 71% still visit the vendor's website after getting an AI recommendation.
• Only 7% of buyers notice a vendor because of brand recognition. Use-case fit is what actually gets you on the shortlist.
• Before August 17, you need to adjust Target CPA/ROAS on budget-constrained campaigns or risk unexpected performance drops.
• Up to 30% of PPC conversions are lost before they ever happen, and Smart Bidding is learning the wrong lessons from that data.
• Content built on first-party data and original research is the only type the synthetic content pool cannot replicate.

B2B buyers start with AI now, not Google

A Semrush survey of more than 600 US B2B professionals found that 84% use AI tools daily and 66% regularly use AI to research vendors. The number that should change how you think about content: only 7% of buyers notice a vendor because of brand recognition. 53% notice a vendor because it precisely matches their use case.

That means a strong brand name will not protect you from a smaller, more specific competitor. AI models respond to concrete prompts ("reporting automation tool for e-commerce agencies with Shopify integration") and if your content does not address those specific scenarios, you simply will not appear. After getting an AI recommendation, 71% of buyers visit the vendor's website and 63% search on Google, so web presence and organic visibility still matter, they just come as the second step, not the first.

The practical implication: stop writing generic brand awareness pages and start mapping the specific use-case combinations your customers actually search for. Charlie's AI agents can accelerate this kind of use-case mapping significantly.

Content gap analysis now takes four hours, not two weeks

Search Engine Land describes a workflow combining Semrush, Google Search Console and an AI model to systematically surface content gaps. The result: analysis that used to take two weeks, done in four hours. Here is how it works:

1. Run a Keyword Gap analysis in Semrush against three to five close competitors. Export keywords with difficulty below 30 and monthly search potential above 500.
2. In Google Search Console, identify queries where you rank in positions 4-10. These have the highest near-term upside, since moving to top 3 typically delivers around 30% more clicks.
3. An AI model scores each opportunity from 0-100 based on volume, competition and relevance, giving you a ranked list instead of a spreadsheet full of raw data.

Average result after three months: a 15% increase in organic impressions and average position moving from 8th to 3rd. These are numbers from a real deployment described in Search Engine Land, not projections.

One important caveat: more than 50% of newly published English content is now AI-generated (Graphite analysis cited in Search Engine Journal). Search systems show a measurable preference for machine-written text because of its statistically predictable style. Accuracy metrics stay stable around 68-70%, so dashboards look green even as source diversity quietly collapses. The content that synthetic pools cannot replicate is first-party data, original research and direct testing. That is your actual competitive moat.

Smart Bidding after August 17: what to do before the rollout

Google clarified in Search Engine Journal what is actually changing: Target CPA and Target ROAS campaigns that are budget-constrained and currently beating their targets will, after August 17, optimize closer to the set target. In plain terms, if your Target CPA is set at $50 but the campaign is actually delivering at $35, after the update it will move toward $50.

What to do before the date:

• Audit all budget-constrained Target CPA/ROAS campaigns and identify where actual performance is significantly better than the set target.
• If you want to keep current performance, lower your Target CPA (or raise your Target ROAS) to match the actual average before August 17.
• After rollout, watch campaigns with the largest gap between actual and target performance most closely. That is where the impact will be most visible.

One thing Google is clear about: budgets will not increase automatically. Budget controls spend. Target CPA/ROAS controls efficiency. If you have been using a constrained budget as an efficiency lever, that approach stops working after this update.

The PPC leaks that Smart Bidding makes worse, not better

The Smart Bidding change would be less critical if the data it learns from were clean. It is not. PPC Hero documents that up to 30% of conversions disappear between click and conversion, in a space that nobody actively owns.

Three of the most common causes:

• GCLID loss during redirects. If a redirect changes the domain or URL structure, the GCLID is dropped and the paid click arrives in GA4 as a direct visit. Your best channel looks like no channel.
• Page speed. Every second of load time costs 7-20% of conversions. Between a third and half of the tags on a typical landing page slow load time by two seconds, which costs more than any headline A/B test will ever recover.
• Spam in CRM forms. Once spam exceeds 10% of form submissions, Smart Bidding actively optimizes for bringing in more of it.

The good news: most of these problems do not require a developer sprint. Edge-layer configuration (typically Cloudflare) handles a large part of it. Start by checking GCLID pass-through across your redirects in Screaming Frog and comparing Google Ads clicks against GA4 sessions for the last 30 days. A gap above 15% signals a tracking chain problem.

What to actually do this week

These four trends form one picture: AI has taken over as the entry point in the B2B buying process, but your website, content and PPC infrastructure need to work cleanly the moment a buyer arrives to verify what AI recommended. Use-case-specific content gets you found. Clean tracking lets Smart Bidding learn the right things. And adjusted bidding targets before August 17 protect the performance you have already earned.

Concrete starting points: run a Keyword Gap analysis in Semrush, check GCLID pass-through, audit Target CPA/ROAS on budget-constrained campaigns, and start planning content built on data only you have. If you want to coordinate all of this across channels without the usual back-and-forth, Charlie is built for exactly that. You can also see how Charlie is priced if you want to run the numbers first.