Foreword
One of the most striking, and surprising, outcomes of our previous Benchmark report of this kind – Fintech Content in the AI Era – was the extent to which analysis of the industry’s content marketing tells the story of the sector’s evolution. Our study of HRTech content didn’t disappoint in that regard either; you can plot the clear transition of content from operational to cultural, peaking in the COVID era as brands sought to retain employee connectivity, and back towards operational and regulatory focus around 2024 as brands emerged from COVID.
In this study, we used our AI to analyse +90,000 pieces of HRTech content across a range of attributes – from keyword optimization to information gain and authority. The goal of the exercise is not to expose specific brands, but to understand the state of content in the space as brands navigate a transition into an era defined by AI-led discovery of information and products.
I was delighted that we were able to publish this latest report alongside our friends at Literal Humans. One of the challenges with spending all day every day thinking about what machines require from content is that it is easy to forget that the ultimate goal is to reach, engage and influence a human being. AI is increasingly an intermediary we must work through – the proverbial middle man or gatekeeper – which adds a layer of complexity but should not obscure the ultimate goal of human influence.
The picture the data paints is clear and reflects the external conditions brands have operated in for more than two decades. Content, almost across the board, is exceptionally well optimized – the category scores 88% overall for accuracy and currency. It is well structured, clear and up-to-date.
But it is woefully undifferentiated.
Just 5.7% of content scores high for information gain and 4.7% demonstrates a strong narrative or unique perspective. That is, the vast majority of content teaches AI things it already knows, and could be lifted from one brand’s website to another’s and nobody would bat an eyelid.
That is a problem that the industry needs to solve, and that solution cannot come from content teams alone. It starts with a mandate from leadership to invest time in quality and originality. My prediction is that the brands that resist the tempting narrative that content is now a “commodity”, and the natural inclination towards hyper-efficiency and mass production that it leads to, will be the ones that succeed. Content may well be a commodity, but perspective and originality is what AI requires to surface that content and that remains – at least for now – a human endeavour.
Key Takeaways for HR Tech Brands
1. Brands must prioritise differentiation and originality.
For two decades, the dominant content question was how do we optimize? First to rank, then to get cited. And broadly speaking, brands got good at it. The category scores 88.1% on accuracy and currency. Structure and formatting are well above threshold.
The AI era demands greater differentiation, though. Does your content offer something that the models don’t already know? The category scores 60.2% on differentiation. Just 5.7% of content shows high information gain. Only 4.7% demonstrates strong narrative ownership.
The new question brands should ask themselves is not “how do we get cited?”. It’s “why should AI cite us?”.
Most HRTech content doesn’t have an answer.
2. AI-assisted production, without a differentiation strategy, makes things worse.
The 2025 content data tells a clear story. Content volume spiked sharply. Optimization scores, already unspectacular, regressed toward the mean. Differentiation fell further still.
This isn’t a warning against AI tools. It’s a warning about how they’re being deployed. Scaling volume without a point of view produces, to be very direct, adequately optimized mediocrity. Brands that adopted AI for velocity, without first asking what they actually have to say, ended up with larger libraries of less distinctive content.
A recent study by Lily Ray, which we covered with our own research too, highlighted the real-world cost of large scale AI production of undifferentiated content, with most brands seeing a brief traffic spike and long term domain damage.
3. Content IP is the proven path. The data shows it working.
Ashby and HiBob both score more than 2x better than the average on depth and differentiation. The common thread is fairly clear in the data: both brands publish things only they can say. Ashby through proprietary recruiting data. HiBob through consistent named research across HR strategy.
There’s a simple model to follow here Identify the unique asset your product or customer base gives you access to – aggregate data, anonymised benchmarks, genuine domain expertise – and build your content strategy around it. The brands winning on differentiation aren’t producing different types of content. They have a deeper knowledge base behind
4. Confident tone, thin substance: your credibility problem with humans.
Based on Literal Humans’ human-centred content analysis.
64% of content in this study is rated high-confidence in tone. Only 6% delivers genuinely high information gain. While our analysis focuses on what AI requires in content, Literal Humans flag the cost in human engagement from this gap.
Readers are good at detecting when a brand is speaking with more certainty than it has earned. They may not articulate it, but they feel it. The brand registers as hollow. Trust erodes quietly. In a category whose entire value proposition is people, 97% of content is functional and efficiency-led. Technically proficient, emotionally inert, and rarely earning the authority it assumes.
The fix isn’t softer language. It’s either backing your claims with evidence, or reserving your most confident voice for content that deserves it.
The State of HRTech
One of the most fascinating, and unexpected, takeaways from this analysis – and our Fintech Content Benchmark before it – was the extent to which we’re able to paint a picture of the category’s evolution through the data. It makes sense; content strategy reflects the buying environment, competitive landscape and marketing strategies of the brands we studied.
So before we get into the detail, we wanted to explore some of the most interesting category-level trends that we were able to pick out in order to contextualise a deeper look at the content landscape specifically.
Transitioning HRTech from Transformation to Infrastructure
Between 2020-2022, the data shows a clear rise in broad workplace transformation narratives: employee engagement, culture, wellbeing, the future of work. HRTech brands positioned themselves as cultural enablers. The glue that held together a happy workforce. Interestingly, this coincides almost perfectly with the COVID pandemic, and HRTech brands clearly reflected the market’s concern around maintaining employee connection and culture as workforces dispersed.
Post-pandemic, we’ve seen a clear shift back towards operational and regulatory content which by 2026, accounts for 71.2% of the category’s output. HRTech brands are repositioning themselves as essential infrastructure for navigating distributed workforces and regulatory complexity. The content evolved to match: more technical, more compliance-heavy and more operationally specific. It has taken on more of a sense-making role.
One interesting feature to a cynic such as myself, when you break the data down based on the primary audience, it’s hard not to question the authenticity of the wellness trend. Across the dataset, just 5.4% of content targeting senior HR executives covers wellness and culture-related topics, vs 32% of content targeting non-senior personas.
Was this ever truly high on businesses agenda?
The Market is Splitting: Platforms vs. Specialists
The category no longer converges around a single content model. There is a visible split between vendors with horizontal ambitions – the “one-stop-shops” – versus those specialising around a specific vertical domain.
This speaks to a maturing category. Earlier stage markets are noisy and often undifferentiated. As they mature, you see some brands compete for market ownership while others carve out defensible niches.
The clearest signal of this is in audience breadth. Contrast two brands: Rippling and Remote.com. Rippling’s content serves five meaningful personas with deep, comprehensive content. This reflects the complexity a platform play imposes on GTM strategy; it serves a complex, multi-stakeholder buying committee. Remote operates a far simpler GTM motion and content operation, capturing the efficiency that focus brings.
Evolution of HRTech Content
The story of HRTech content over the last decade is a story of systematic professionalisation as brands built infrastructure and processes for consistent, well structured search-era content. The data also shows clearly how AI disrupted those processes, for better and for worse, and the extent to which brands are unprepared for the requirements of the AI era.
Highly-tuned Content Machines
HRTech content teams, by and large, operate extremely efficiently. The current model is structured, commercially integrated and mid-funnel dominant.
MOFU content significantly outweighs TOFU and BOFU content. “Guide” has become the dominant format over blog or article, reflecting the sensemaking role that we mentioned previously.
All of this speaks to increasingly commercially aware content strategies, focused on serving a critical role within a multi-stakeholder and complex buyer journey. Content has evolved past pure awareness-generation and taken on more of a buyer enablement and education role.
Opportunity for Technical and Editorial Innovation
Across the sample, interactive content is extremely uncommon – just 1.3%. External contributors and thought leadership is also declining in recent content. We saw a similar trend in our Fintech study; brands are favouring velocity and volume over creativity and innovation.
We believe we’ll see more innovation. AI rewards clear, demonstrated authority, so thought leadership is critical. We also believe that as more of the best content is ungated, we’ll see more innovation in content formats and storytelling as brands work harder to capture human attention and provide onsite value beyond what users can access through AI summaries of their content. Innovative brands, armed with Claude code and other vibe coding tools, can be much more creative with how “content” is defined and packaged to engage readers.
Content is well optimized, poorly differentiated.
One striking finding in the data is the stark gap between content optimization (structure, extractability, keyword balance) and differentiation (information gain, narrative ownership, evidence depth).
Interestingly, the improvements in content optimization halted quite sharply in 2025. While optimization remains solid – certainly compared with differentiation – there is a clear regression to the mean. We wanted to investigate, and it turns out 2025 also happens to have seen an explosion in content volume and velocity.
The takeaway seems clear:
- Content has always been far better optimized than differentiated.
- In 2025, brands aggressively adopted AI-driven content, leading to a huge uptick in content volume.
- The result was a regression to the mean. Content didn’t become badly optimized, it became average. The hallmark of AI content.
- Differentiation fell even further.
This is the most important finding of this study: content is highly optimized, but extremely poorly differentiated. While AI-generated content has had a damaging effect on both, it’s had a disastrous effect on differentiation.
The big problem brands face is that in the AI era, differentiation is what drives success. Google’s latest core update clearly rewarded information gain (content featuring unique original research) with a +22% visibility increase.
Search demands clarity on a given topic, typically chosen for its intent or volume. The goal is to provide the best possible answer to a given question, in order to “rank”.
AI demands differentiation. Buyers can generate answers tailored to their question and circumstance in seconds, raising the bar for brands. To be useful to Humans or AI you must provide something they can’t get elsewhere.
For 20 years, brands have asked the question “how do we optimize our site?”. First to “rank” and now to “get cited”. And generally speaking, they got pretty good at that. Very few have stopped to ask the more important question: “Why should AI cite us?”.
In the next section, we take a deeper look at the challenge ahead.
Content Quality & AI Era Readiness
So, how ready is HRTech content for the AI era? Analysing 90,000 pieces of content across 5 dimensions, comprised of +20 specific attributes, the category scores 73.8% overall. That number masks a lot of variation, though, between the dimensions brands have historically optimised for and those they have not.
| Dimension | Score | Verdict |
|---|---|---|
| Accuracy & Currency | 88.1% | Strong. 86.3% of content is current. Content maintenance is a genuine category strength. |
| Content Quality | 78.5% | Solid. Structure and formatting are well above threshold. Comprehensiveness is weaker: only 25.6% qualifies as “strong”. |
| AI Extractability | 73.3% | Competent. Well-structured for machine parsing. |
| Citation Likelihood | 68.8% | Below par. Only 7.6% has high authority signals. Source attribution is a specific weakness. |
| Strategic Differentiation | 60.2% | The gap. The weakest dimension by a significant margin. Low information gain and narrative ownership throughout. |
Where the Category is Strong: Accuracy, Currency, Quality.
HRTech content is overwhelmingly accurate, up-to-date, and generally of a high quality. Only 0.2% of HRTech content is over-optimised. 92.6% is adequately or well structured. 88.1% of content scores strongly on accuracy and currency.
Compared with our similar study of Fintech content, HRTech brands devote a far greater time to content maintenance. There is a sharp uptick in 2023, which appears to show HRTech brands leveraging AI for content maintenance and improvement in a way that Fintech brands have not.
This likely reflects the focus on operational and regulatory sensemaking – currency is not a question of discipline, it’s critical to the content’s purpose. This is positive; AI forms perspectives based on all your content not just your most recent content, so cultivating an overall content footprint that is high quality, clear, consistent and current is important.
The main area for improvement is in comprehensiveness. While content is high quality and structured, it does not always provide the depth that is required in order to produce more meaningful analysis in the AI era.
AI Extractability: Content is Readable, but not Citable
We often push back against citations as a “North Star” of AEO. Our Dark AI research showed that within a complex buyer journey citations are rare, occurring in just 16% of AI responses. As a metric, citations miss too much of the impact, challenge and opportunity AI presents.
That doesn’t diminish their value, though, and while AI extractability has improved measurably across the dataset there is a gap between content AI can parse and content it would cite.
Just 27.8% of content achieves strong extractability, and just 7.6% signals authority clearly. Brands are structuring their content for AI, but they’re not giving it a reason to cite it. Very little content excels in any of the aspects that make it easy for AI to directly cite content.
The Big Gap: Strategic Differentiation
At this point, it is useful to summarise the framework we’ve developed at Demand-Genius for measuring information gain. You can see a full overview here, but we identify three tiers of information gain:
- Level 1 – Interpretive Gain: A new perspective on existing knowledge.
- Level 2 – Empirical Gain: Original research like surveys, data, or anything else that measurably expands knowledge.
- Level 3 – Conceptual Gain: A genuinely novel concept, framework or mental model that is substantiated by data or expert analysis.
You can use our Information Gain Calculator to assess any content or content brief against this framework, or evaluate your brand’s content at scale through Demand-Genius.
We analysed all content against this framework, as well as assessing narrative ownership as a broader means of assessing whether that content is differentiated.
The data is alarming.
- 4.7% of content displays strong narrative ownership.
- 5.7% shows high information gain.
We discussed the trend earlier, but this data shows that regardless of AI’s impact, the overwhelming majority of HRTech content is undifferentiated. That content will be synthesised by AI but it will not exert any influence on the way it presents your category or brand, and it does not give it any reason to surface it explicitly.
An interesting trend emerges when we dig a bit deeper into the data to see where we do see higher originality. Brands and content with access to proprietary data, or more vertically-focused brands with deep domain expertise, consistently produce higher quality content.
Note : Data-led content is content where the primary evidence basis is quantitative. Vertically-specific content is content covering specialist topics such as global mobility or international expansion.
There’s a really clear strategic roadmap here for Marketing teams. It’s important brands think again about that question: why should AI cite us? Building some form of Content IP, ideally on a unique advantage your product affords such as aggregate datasets or customer access, enables more differentiated AI-era content strategies.
Two Case Studies: Ashby and HiBob
What stronger AI-era content looks like: Ashby and HiBob
We wanted to look at two vendors in the dataset that stand out as stronger than the average in terms of AI-era readiness, and what that readiness is built on. Those brands are Ashby and HiBob, both of whom perform more than 2x better against our content depth and differentiation measures.
Ashby’s advantage is derived from a strong content IP – proprietary recruiting data – and a category point of view built on things Ashby’s data uniquely positions them to quantify.
HiBob shows that a large content operation and genuine authority are not mutually exclusive. The combination of breadth and research (guides, case studies, product content, webinars, and named research leadership across HR strategy topics) creates a content library that consistently surfaces proprietary insight.
This is another really valuable takeaway for brands. The common denominator between the two highest performers is not format or volume. It is the depth of the knowledge base behind the content. Both brands are publishing things that only they can say.
Content That Moves People
Written by Paul David, CEO at Literal Humans
The team at Demand-Genius have given us a forensic picture of HRTech content from the outside in: what machines can parse, extract, and cite. Our job at Literal Humans is to look at the same content through the lens of the people ultimately on the receiving end of it.
And when we do that, one number keeps stopping us in our tracks.
97% of HRTech content is classified as functional and efficiency-led.
Practically none of it operates in emotional or transformational territory. For a category whose entire value proposition is people (how you hire them, retain them, develop them, motivate them) that’s a striking finding.
It also isn’t surprising. Functional content is easier to produce, easier to justify, and easier to optimise for search. Efficiency narratives have a clear buyer job-to-be-done. So brands gravitated there, stayed there, and the whole category moved in the same direction together. The result is a content landscape that is technically proficient and emotionally inert.
The confidence problem
There’s a second finding that deserves attention alongside the 97% figure, because the two together tell a more complete story.
64% of content in this study is rated “high confidence” in tone. Only 6% delivers genuinely high information gain.
That gap (confident voice, thin substance) is worth sitting with. These aren’t poorly written pages. They’re competent, structured, and pass every technical benchmark. But they’re making claims with a certainty that the underlying content doesn’t earn.
From a human psychology standpoint, this matters enormously. Readers are exceptionally good at detecting the mismatch between tone and substance. They might not be able to articulate it, but they feel it. The brand registers as hollow. Trust erodes quietly, invisibly, and often irreversibly.
This isn’t a new problem. But AI has made it urgent in a way it wasn’t before.
Why human-centred content is now a competitive advantage
For years, the volume of competent-but-undifferentiated content was manageable. Buyers filtered it, search surfaced the best-optimised version of it, and brands with bigger content operations generally won. The system had friction, but it worked well enough.
AI removes most of that friction, and with it, most of the cover.
When a buyer can generate a comprehensive, personalised answer to any procurement question in seconds, the bar for what earns their attention shifts fundamentally. They’re no longer comparing your content to your competitors’ content. They’re comparing it to a direct answer to their exact question, tailored to their context, with no filler.
In that environment, functional and efficiency-led content (however well-optimised) is essentially commoditised from day one. AI can synthesise it. It cannot replicate content that carries a genuine human perspective: a point of view that someone earned, a story that only that brand can tell, a piece of insight that came from living inside the problem rather than describing it from a safe distance.
This is what we mean by human-centred content. Not soft content. Not content that prioritises feeling over substance. The opposite, in fact: content with the confidence to say something specific, with the evidence to back it up, and with enough editorial courage to actually take a position.
What the best content does differently
The brands that stand out in this dataset (as Tom has highlighted with Ashby and HiBob) share a common characteristic. They publish things that only they can say.
That specificity is itself an emotional signal. It tells a reader: someone thought carefully about this. Someone has skin in the game. That feeling (hard to manufacture, impossible to automate) is the foundation of trust in a buying relationship.
The practical implication for HRTech marketing teams isn’t to abandon functional content. It’s to stop treating functional content as the ceiling. The 97% figure represents a category that is systematically underinvesting in the content that builds long-term brand equity, earns genuine AI citation, and (most importantly) actually moves the people who eventually make buying decisions.
Buyers are human. Humans respond to content that treats them as such. In a category that exists to serve the human side of business, that shouldn’t be a radical proposition.
