Originally published · Updated
[ TL;DR ]
Key takeaways
- AI didn’t fix the expertise gap. Agencies can now produce technically plausible content at scale, but generating a convincing explanation is different from understanding, validating, and adding to it.
- GEO risks becoming the new keyword stuffing. Mass-producing repetitive answers, comparison pages, and FAQs to influence AI responses repeats the old SEO mistake: optimizing for the distribution mechanism, not the reader.
- AI citations need context. Mentions and citations are visibility indicators, not business outcomes. Measure accuracy, buyer relevance, and demand alongside them.
- Faster drafts don’t always mean lower costs. If technical research is weak, your team still pays for validation, corrections, and reviews, often on more content than before.
- Evaluate agencies on five capabilities: technical expertise, AI governance, credible GEO/AEO strategy, meaningful measurement, and the ability to reduce your team’s workload.
Tech marketing agencies still fail at content because many lack the technical expertise needed to produce accurate, differentiated, and valuable material, even when AI makes content creation faster and cheaper.
AI has changed almost everything about how marketing agencies produce content. The tools are faster, the costs are lower, and the volume of content that an agency can deliver has increased dramatically.
But there’s one problem AI hasn’t solved: the lack of genuine technical expertise.
Back in 2019, I wrote about why so many technology companies were disappointed with their marketing agencies. The underlying issue was simple. Agencies knew how to market, but they often didn’t understand the technologies they were supposed to be marketing.
Today, I see the same problem taking a different shape. Instead of relying on generalist writers to produce superficial articles, agencies can use generative AI to produce large volumes of technically plausible content.
And now, with the growing interest in Generative Engine Optimization (GEO) and Answer Engine Optimization (AEO), we’re seeing another troubling development: agencies optimizing content to influence AI-generated answers, sometimes without asking whether that content offers anything valuable to the actual reader.
We may be witnessing the next generation of keyword stuffing, only this time it’s designed for AI engines rather than traditional search rankings.
That doesn’t mean GEO or AEO are inherently bad practices. Quite the opposite. AI search is changing how buyers discover, research, and compare technologies. Brands need to adapt.
But the fundamentals of good technical content haven’t changed. In fact, I would argue they’re becoming more important.
1. Why Have Tech Marketing Agencies Historically Struggled With Technical Content?
Tech marketing agencies have historically struggled with technical content because their marketing capabilities often exceed their understanding of the products and technologies they’re promoting.
What Do Failed Agency Engagements Reveal?
Failed agency engagements often reveal a gap between the marketing services being delivered and the technical expertise needed to make those services effective.
Let me start with two stories from the original version of this article.
In one conversation, the VP of marketing and sales at an established big data company told me about a pilot campaign with a leading marketing agency. The campaign cost $14,000 and generated two leads, effectively $7,000 per lead.
In another case, a technology startup had been spending approximately $150,000 annually with an inbound marketing agency. The team liked the people they worked with, but they were deeply disappointed with the content.
When I asked about their performance, they couldn’t initially tell me their monthly traffic. After looking into it, they reported approximately 3,000 monthly visitors.
These examples were never meant to suggest that traffic or cost per lead alone can determine content ROI. Enterprise buying journeys are complicated, and not every content investment produces immediate measurable demand.
But they raised an important question: What exactly were these companies paying their agencies to deliver?
Why Does a Lack of Expertise Create More Work for Clients?
A lack of technical expertise creates more work because internal teams must repeatedly explain products, correct inaccuracies, and compensate for knowledge gaps.
The problem wasn’t necessarily that the agencies lacked talented marketers. They lacked the technical expertise to understand the products, engage with the target audience, and develop meaningful content.
A typical marketing agency could manage campaigns, advertising, social media, email marketing, and SEO. But when it came to explaining the architectural trade-offs of a cloud infrastructure solution or the technical differences between cybersecurity products, the agency frequently had to depend on its client’s internal experts.
This created an expensive and inefficient process.
Marketing teams spent their time translating technical information for external writers. Engineers and product managers were repeatedly asked to explain basic concepts. Content drafts came back with inaccuracies or superficial explanations.
The agency was supposed to reduce the company’s workload. Instead, it created more work.
Today, AI can accelerate this entire process. But accelerating a flawed process doesn’t necessarily fix it.
2. Why Can’t AI Replace Technical Expertise in Content Production?
AI cannot replace technical expertise because generating a convincing explanation is different from understanding, validating, and contributing original knowledge about a technology.
What Has AI Actually Changed About Content Production?
AI has made research preparation, outlining, drafting, and editing substantially faster, reducing many of the traditional costs associated with producing content.
The arrival of generative AI has fundamentally changed the economics of content production.
An agency no longer needs to spend days researching and drafting a basic technical article. AI can produce outlines, summarize documentation, identify related questions, generate comparisons, and draft long-form content in minutes.
This is valuable. At IOD, we see tremendous opportunities for AI to support technical experts, writers, editors, and marketing teams.
The problem begins when agencies confuse the ability to generate an article with the ability to understand its subject.
Why Does Technical Judgment Still Matter?
Technical judgment matters because someone must evaluate the accuracy, relevance, and originality of the information AI produces.
Consider a cybersecurity company developing content about cloud detection and response.
An AI model can explain the fundamentals, summarize common threats, and generate a convincing comparison of competing solutions. But can the person supervising the process evaluate whether the explanation is technically correct? Can they challenge a product claim? Can they distinguish a real architectural advantage from generic vendor messaging?
More importantly, can they contribute anything new?
Technical buyers don’t need another article explaining that cloud security is important. They want to understand how a solution works, where it fits into their infrastructure, what compromises it introduces, and how it compares with alternatives.
These answers require subject-matter knowledge, practical experience, research, and judgment.
AI is extremely useful when the people directing it have that knowledge. Without it, agencies risk producing content that looks professional but contributes little.
The real opportunity is to combine AI-assisted production with experienced technical professionals who can introduce original perspectives, validate claims, and challenge assumptions.
This is very different from using AI to replace expertise.
3. Is GEO Becoming the New Keyword Stuffing?
GEO risks becoming the new keyword stuffing when agencies mass-produce repetitive answers, comparison pages, and FAQs primarily to influence AI-generated responses rather than help buyers make informed decisions.
How Is GEO Repeating the Mistakes of Traditional SEO?
GEO is repeating traditional SEO mistakes when agencies prioritize the number of optimized pages and answer variations over the originality and usefulness of their content.
For years, traditional SEO encouraged marketers to produce content around specific keywords and search queries.
The approach wasn’t inherently wrong. Search demand remains a useful way to understand what people want to know.
But it also encouraged some agencies to produce large quantities of shallow content designed primarily to capture rankings.
We all remember the results: repetitive listicles, generic how-to articles, superficial comparisons, and endless variations of the same subject.
Now, a similar pattern is emerging around GEO and AEO.
Instead of targeting only a search keyword, marketers are targeting the questions buyers might ask ChatGPT, Google AI Mode, Perplexity, or other AI-powered discovery tools.
For example, an agency working with a cloud security vendor might identify questions such as:
- What are the best cloud security platforms for enterprises?
- Which CNAPP solution is best for a multi-cloud environment?
- What are the leading alternatives to a particular cloud security vendor?
- Which cloud security platform offers the strongest runtime protection?
These are legitimate buyer questions, and answering them well can create value.
The problem is what happens next.
An agency identifies hundreds of similar questions and uses AI to generate individual articles, FAQs, comparison pages, and other assets around each variation.
The content may be grammatically correct and even technically plausible. But much of it repeats the same information, offers little original analysis, and exists primarily to increase the likelihood of the brand appearing in AI-generated answers.
This is what I call the new keyword stuffing: replacing keyword repetition with large-scale answer and topic repetition.
It’s an editorial description of the practice, not an official Google penalty category.
The objective has changed from appearing in traditional search rankings to influencing AI-generated responses. But the underlying mistake is familiar: optimizing for the distribution mechanism rather than the information needs of the audience.
What Separates Legitimate GEO From GEO Stuffing?
Legitimate GEO improves the accessibility, accuracy, and usefulness of existing knowledge, while GEO stuffing manufactures repetitive content primarily to increase AI visibility.
Good GEO helps AI systems discover, understand, and accurately reference useful information. It involves clear technical explanations, accessible content, original research, authoritative sources, and consistent product information.
Bad GEO tries to manufacture visibility by multiplying generic content, creating artificial mentions, or publishing slight variations of material already available everywhere else.
Google’s guidance on optimizing for generative AI search reinforces the importance of original, valuable content and established SEO fundamentals rather than tricks intended to influence AI-generated answers.
Google’s spam policies also address scaled content abuse: generating large volumes of unoriginal content primarily to manipulate search rankings without helping users.
Importantly, this doesn’t mean AI-generated content is automatically penalized. It means the purpose, quality, and value of the content matter.
The lesson for B2B technology brands is clear.
Don’t confuse a GEO content strategy with a plan to publish hundreds of pages answering slightly different versions of the same question.
4. Why Can AI-Generated Technical Content Sound Right but Still Be Wrong?
AI-generated technical content can sound credible while containing errors because language fluency does not guarantee technical accuracy, current product knowledge, or correct interpretation of evidence.
What Makes Technical Inaccuracies So Difficult to Detect?
Technical inaccuracies can be difficult to detect because AI-generated explanations often sound authoritative even when they misrepresent architectural concepts, product capabilities, or technical evidence.
One of the most frustrating problems I described in the original article involved a company whose marketing agency claimed to have a qualified data expert.
The agency produced an article about time-series data that contained so many technical inaccuracies the client abandoned it.
This problem hasn’t disappeared.
In fact, generative AI introduces a particularly challenging version of it.
AI-generated technical content can be convincing even when it contains incorrect architectural assumptions, outdated information, unsupported benchmarks, or inaccurate product claims.
An inexperienced writer may not recognize these problems because the article reads well.
The risk goes beyond publishing a factual mistake.
For a cybersecurity, cloud infrastructure, or data platform vendor, technical accuracy is part of the brand’s credibility.
A developer evaluating an integration, an architect comparing platforms, or a security engineer researching deployment options will notice when an explanation doesn’t reflect how the technology actually works.
And once credibility is damaged, the consequences can extend beyond a single blog post.
How Can Agencies Ensure Technical Accuracy?
Agencies can improve technical accuracy by involving qualified subject-matter experts throughout research, planning, drafting, and validation rather than relying on final-stage proofreading.
This is particularly visible in cybersecurity marketing, where agencies must create content for practitioners, technical decision-makers, and business buyers with very different expectations.
In our comparison of seven cybersecurity content agencies for 2027, we examine practitioner credentials, technical validation, production costs, GEO/AEO capabilities, and audience fit.
One of the key distinctions is between an agency that can demonstrate genuine technical expertise and one that can merely produce professionally written cybersecurity content.
This is why I believe technical content production needs an explicit validation process.
Subject-matter experts should participate in the research, planning, and content development stages, not simply receive a finished AI-generated article for approval.
They should be able to challenge the angle, validate evidence, add original insights, and identify areas where the content misrepresents the technology.
The AI model should support this process, not determine what is technically true.
5. Are AI Citations and Brand Mentions Reliable Measures of Content Success?
AI citations and brand mentions are useful visibility indicators, but they are not reliable measures of business success unless interpreted alongside relevance, accuracy, buyer engagement, and commercial outcomes.
Why Is AI Visibility Different From Traditional Organic Traffic?
AI visibility differs from traditional organic traffic because users can discover and evaluate brands directly within AI-generated answers without necessarily visiting the cited websites.
Traditional SEO focused heavily on keyword rankings, impressions, organic traffic, and conversions.
GEO introduces additional metrics, including AI citations, brand mentions, visibility across prompts, and AI share of voice.
These can be useful indicators. But they are not business outcomes by themselves.
A brand might appear frequently in AI-generated answers without seeing meaningful growth in qualified opportunities. Conversely, a technically authoritative article may influence buying decisions without generating substantial traffic.
This distinction matters because AI search is changing user behavior.
A July 2025 Pew Research Center study found that users clicked traditional search-result links in 8% of visits where an AI summary appeared, compared with 15% of visits without one. The research examined Google search behavior among a sample of U.S. adults, so these figures shouldn’t be treated as B2B-specific conversion benchmarks.
Nevertheless, they demonstrate an important shift: visibility and website traffic are becoming less directly connected.
How Should Agencies Measure GEO Performance?
Agencies should measure GEO performance by combining AI citations, brand mentions, answer accuracy, buyer relevance, and available demand-generation indicators instead of focusing on visibility metrics alone.
The temptation for agencies is to replace one set of vanity metrics with another.
Instead of reporting that an article ranks for 50 keywords, an agency may report that the brand appears in 100 AI-generated responses.
But what does that actually tell the marketing team?
Was the product recommended accurately? Was it associated with the right technical capabilities? Did it appear in relevant buyer scenarios? Was the mention favorable, neutral, or based on inaccurate information?
And ultimately, is the brand reaching the audience it wants to influence?
For B2B technology companies, GEO measurement should connect AI visibility with accurate positioning, meaningful audience engagement, branded demand, and qualified pipeline wherever attribution is possible.
AI citations and mentions matter. But they need context.
Being mentioned by an AI system is not the same as being recommended by a knowledgeable technical buyer.
6. Does Faster AI Content Production Really Reduce Marketing Costs?
Faster AI content production reduces drafting costs, but it doesn’t necessarily reduce the total cost of content when technical validation, internal reviews, corrections, and coordination remain expensive.
Which Content Production Costs Does AI Actually Reduce?
AI primarily reduces the time and cost of repetitive research preparation, outlining, drafting, and editing tasks, but its impact depends on the quality of the surrounding workflow.
One of the strongest arguments for AI-powered content production is efficiency.
And rightly so.
When properly implemented, AI can dramatically reduce time spent on research preparation, outlining, repetitive drafting tasks, and editorial coordination.
However, there is an important difference between reducing production effort and reducing the total cost of producing a valuable content asset.
Think about the traditional agency workflow.
A marketing manager prepares a brief. A generalist writer drafts the article. The marketing manager reviews it. A technical expert identifies inaccuracies. The article goes back for revisions. Another review follows.
Now replace the writer’s initial drafting work with AI.
The draft arrives sooner. But if the underlying technical research is weak, the client still needs to identify the inaccuracies, clarify the product positioning, provide missing information, and manage the revision process.
Why Can AI Content Increase Internal Review Costs?
AI content can increase internal review costs when faster drafting produces more material than internal experts can efficiently evaluate, correct, and approve.
In some cases, the process may become even more frustrating because the team receives more content than it can realistically validate.
The agency may report improved production efficiency while the client’s internal review workload increases.
This is why the economics of content production should be evaluated beyond price per article or words produced.
A more meaningful calculation considers the full cost of research, technical validation, internal review, revisions, coordination, and the content’s eventual contribution to business goals.
The most effective AI-enabled workflows do not simply generate drafts faster.
They reduce unnecessary work across the entire production process by combining reliable technical knowledge, expert participation, reusable brand context, and structured review.
7. How Should B2B Tech Brands Evaluate a Content Agency for 2027?
B2B tech brands should evaluate content agencies based on five capabilities: technical expertise, AI governance, credible GEO/AEO strategy, meaningful performance measurement, and the ability to reduce internal workload.
These criteria are particularly important when evaluating agencies that promise high-volume AI content production.
Can the Agency Demonstrate Technical Expertise and AI Governance?
A capable agency should combine identifiable technical experts with a controlled AI production process that ensures accuracy, reliable sourcing, and consistent product messaging.
Ask who will actually work on your content.
Don’t settle for a general statement about access to technical writers or AI-powered research tools. Ask to meet the people responsible for the technical substance.
A capable technical expert should be able to understand your product quickly, challenge your messaging, identify valuable topics, and contribute perspectives that your internal team hasn’t necessarily considered.
Agencies should also be using AI to improve efficiency. Refusing to adopt the technology is not a competitive advantage.
But ask how AI is incorporated into their process.
What sources does it rely on? How are product claims validated? Who reviews technical explanations? How is conflicting information handled? How do they keep content aligned with evolving product capabilities and messaging?
The answer should demonstrate an actual production methodology, not simply a collection of AI tools.
Can the Agency Deliver GEO Results Without Increasing Your Workload?
An effective agency should build GEO visibility through authoritative content, measure meaningful outcomes, and reduce the amount of research, validation, and revision work required from the client’s team.
Ask how the agency plans to improve your visibility in AI-powered search and discovery.
If the primary recommendation is to generate hundreds of keyword variations, FAQs, or comparison pages, be cautious.
A stronger strategy should identify the information your buyers need, the subjects on which your company has genuine expertise, and the original evidence you can contribute.
For many B2B technology brands, this may include technical tutorials, architecture guides, practitioner insights, benchmarks, research reports, and evidence-based comparisons.
Not every asset needs to be long or complex. But every asset should have a clear purpose and add something meaningful.
Agencies should also understand how to monitor organic search performance alongside AI visibility.
Ask how they distinguish brand mentions from citations, monitor accuracy, track changes over time, and connect content performance to meaningful demand indicators.
Most importantly, ask how insights from this measurement change the editorial strategy.
Finally, evaluate how much work your team will need to perform.
An agency that requires your product managers and engineers to rewrite articles isn’t solving your content problem.
The right partner should be able to operate with a high degree of technical independence while collaborating with internal experts at the moments when their input is most valuable.
That’s where AI and technical expertise can become a particularly powerful combination.
For a more detailed assessment, see our guide on how to evaluate a cybersecurity content agency for 2027. It provides a practical framework for evaluating technical depth, access to subject-matter experts, AI governance, and managed delivery responsibilities.
8. Why Is Expert-Led Content Still Critical to B2B Tech Marketing?
Expert-led content remains critical because technical buyers need accurate information, original insights, and credible perspectives that neither generic writers nor AI-generated content can consistently provide on their own.
Why Does Original Expertise Matter More in an AI-Driven Market?
Original expertise matters more because AI makes generic information widely accessible, increasing the value of credible technical insights, firsthand knowledge, and differentiated perspectives.
When I originally wrote this article in 2019, my message was that technical content is too important to leave entirely in the hands of marketers who don’t understand the technology.
Seven years later, I believe that argument is even stronger.
The content ecosystem is changing. AI assistants are becoming an important part of how technology buyers research solutions. GEO and AEO are introducing new opportunities and challenges. Production costs are falling, and agencies can publish at a scale that previously would have been impossible.
But content volume is not a substitute for expertise.
The brands that benefit most from these changes won’t necessarily be those publishing the largest number of AI-optimized articles.
They will be the brands able to communicate their expertise clearly, contribute original knowledge, support their claims with evidence, and build credibility with both technical buyers and the systems those buyers use to discover information.
What Should Your Technology Marketing Agency Ultimately Deliver?
Your technology marketing agency should transform your company’s knowledge into credible content that helps buyers understand your products, strengthens your positioning, and supports long-term growth.
This requires more than a content calendar or an AI writing tool.
It requires technical professionals, strong editorial judgment, a reliable body of product and brand knowledge, and a production process that uses AI to amplify human expertise.
Your technology marketing agency should help you turn what your company knows into something the market can learn from, trust, and act on.
That was the point in 2019. It’s still the point in 2026.
Only now, the cost of getting it wrong may be much higher.
At IOD, we combine technical expertise, human-led editorial production, and AI-powered workflows to help B2B technology brands create content that serves their buyers, strengthens their positioning, and supports long-term growth.
[ FAQ ]
Frequently Asked Questions
Why do tech marketing agencies still fail at content in the age of AI?
Tech marketing agencies still fail at content because many lack the technical expertise needed to produce accurate, differentiated, and valuable material, even when AI makes content creation faster and cheaper. AI cannot replace technical expertise because generating a convincing explanation is different from understanding, validating, and contributing original knowledge about a technology.
Is GEO becoming the new keyword stuffing?
GEO risks becoming the new keyword stuffing when agencies mass-produce repetitive answers, comparison pages, and FAQs primarily to influence AI-generated responses rather than help buyers make informed decisions. Legitimate GEO improves the accessibility, accuracy, and usefulness of existing knowledge, while GEO stuffing manufactures repetitive content primarily to increase AI visibility.
Are AI citations and brand mentions reliable measures of content success?
AI citations and brand mentions are useful visibility indicators, but they are not reliable measures of business success unless interpreted alongside relevance, accuracy, buyer engagement, and commercial outcomes. Being mentioned by an AI system is not the same as being recommended by a knowledgeable technical buyer.
How should B2B tech brands evaluate a content agency for 2027?
B2B tech brands should evaluate content agencies based on five capabilities: technical expertise, AI governance, credible GEO/AEO strategy, meaningful performance measurement, and the ability to reduce internal workload. A capable agency should combine identifiable technical experts with a controlled AI production process that ensures accuracy, reliable sourcing, and consistent product messaging.
