Revolution When: AI and the Open Web

Over the past 30 years, white collar work has become computer work. An increasing number of careers have moved online, creating new jobs for those that want them. AI threatens that dynamic. Already we're seeing downsizing in tech firms, although this is related more to the fragile state of the economy than anything else. More fundamental changes could happen, but that requires realism from all parties about the current state of the tech industry. Ultimately, the tech industry is not the growth sector it once was. The SaaS revolution drove a wave of digital transformation that has run its course.

The trajectory of Salesforce is emblematic of this trend. Just a year ago, Marc Benioff was under immense pressure to cut R&D and slow product releases. Since launching Agentforce, he has become the poster boy for agentic AI. That's not a coincidence. From a Wall Street perspective, Agentforce is not so much about the future, but more about turning back the clock. Tech firms want to launch the next wave of digital transformation. AI is a chance to boost margins and recapture the double or triple digit growth rates of the past. Business leaders are onboard with this, but are still struggling with the maturity of LLM technology.

Disrupting Marketers

Generative AI investment is still having real world consequences though, particularly within the digital marketing ecosystem. Enterprise marketing is increasingly being done in-house. Internal teams are being used for media and content work that historically was outsourced. We're already seeing the consequences within several niches of the digital marketing ecosystem such as copywriters, translators, and SEO specialists. Marketing agencies are struggling. It's worth re-iterating that AI is not the primary driver of this trend. Slow growth and tightening marketing budgets are more significant factors. However, AI generated content makes in-sourcing easier, particularly within enterprise.

The same struggles are affecting media publishers, an industry mostly funded by marketers. Genuine concerns are being raised about the viability of ad-funded media, given steep declines in web traffic. That's not a new debate, although ChatGPT has given it a sharper impetus. Journalism has been struggling for a long time because digital media has rarely been profitable for publishers. The news media does have one key advantage though: government backing. We've already seen various laws requiring Google to pay royalties when linking to news articles. The importance of Google News to publishers has hampered the effectiveness of these laws. However, if Google stops driving traffic to newspaper websites, then their owners will have less reason to co-operate with search engines.

Newspapers currently earn more from Google News referral traffic than from the royalties enforced in certain jurisdictions. That balance is beginning to change, which in turn, will seriously harm the accuracy of LLMs, particularly when it comes to breaking news. Publishers are already erecting paywalls for website scrapers. Meanwhile, Cloudflare are developing technology to enforce those paywalls. Despite best efforts, there's no way around those paywalls. Perplexity have been caught trying to bypass them, which will have real consequences if that behaviour continues. Publishers, marketers, and search engines all have a symbiotic relationship. AI doesn't change that. It merely changes the business model.

A Symbiotic Relationship

Tech firms are being forward thinking about the user experience. Enterprises are trying to fix their data in an effort to catch up. Meanwhile, the real economy is left behind. In part, because tech firms are yet to work out the business model for AI. Relying on old playbooks isn't working. Generative AI is directly threatening the profitability of big tech's most important ad customers. In the short term, that will cause challenges to marketers and publishers. In the long term, though, it's tech firms that will face the consequences. Paywalls on training data are part of the reason, but there is a far more fundamental threat to the profitability of AI search.

The likes of Google and OpenAI are betting on advertising to fund future model development. However, it's not clear how this will work in an AI search interface. It's a question of volume. AI generates fewer clicks than traditional search. Sure, conversion rates are higher, but not sufficiently high to outweigh reduced click volumes. Advertising ROI will go down — which in turn will reduce the number of PPC ads being placed. That's annoying for marketers, who have alternative advertising channels. Google don't have alternative revenue streams. Subscriptions certainly won't cover the difference. Consumers simply won't pay.

AI vendors need to map out a solution to this dilemma. Yet they're too busy scrapping data to consider the needs of digital marketers or the open web. The relationship between AI and publishers needs to be resolved sooner rather than later, otherwise governments and the market will force a detrimental solution on tech firms. Shifting the cost burden onto business customers won't work. Businesses see the potential, but aren't seeing the outcomes. That's not an issue at the moment. AI development is funded by either venture capital or monopoly profits. Neither of these are sustainable in the long term. VC firms are willing to pay for AI right now. At some point, they'll want a return. Currently, it's not clear where that return will come from.