So how does AI in thought leadership benefit the audience, exactly?

Successful integration of AI in thought leadership will depend on adopting an audience-first approach, rather than focusing on the low hanging fruit of cost savings and productivity gains.

One of my favourite business book titles is Where are the Customers’ Yachts? by Fred Schwed. Written in 1940 about the world of Wall Street investing, it starts with a story of a visitor to New York who is admiring some expensive boats in a marina and told that they all belong to bankers and brokers. The visitor then asks the question that forms the title of the book – only to be told that, despite following the advice of their bankers, the customers can’t afford them.

I was thinking about this title recently when I was reviewing the rapid development of AI in thought leadership. According to McKinsey’s State of AI in 2023, 14% of sales and marketing functions are using generative AI. Of course, this is only just the beginning: a report from Europol forecasts that 90% of online content could be AI-generated by 2026.

For the most part, the emphasis of all this activity is on productivity and efficiency, rather than on how it could benefit the customer, or audience. A recent JournalismAI survey from LSE, Polis and Google News Initiative found that more than half of newsrooms cited increasing efficiency and productivity as key goals for AI adoption. They hope to automate monotonous tasks, streamline workflows and allow journalists to engage in “more creative, relevant, and innovative work”.

It is easy to understand why media organisations and content marketers are rushing to make use of AI in this way. Budgets are tight and resources are scarce, and any tool that helps to simplify content production is sure to be welcomed. But in all the excitement about the potential of AI to make their jobs easier and provide quick ROI, there is one important stakeholder whose views ought to be paramount: a bit like the visitor to New York asking after the customers’ yachts, my question is: ‘What does AI actually do to benefit the audience?’

Now, there is nothing wrong with using technology to introduce more efficiencies – but I am not sure that the bottom line ought to be the starting point

Now, there is nothing wrong with using technology to introduce more efficiencies – but I am not sure that the bottom line ought to be the starting point. My sense is that content marketers who think ‘audience first’ will generate better, more sustainable value from Gen AI than those that focus primarily on efficiency. If we think only about efficiency and automation to be more productive and save costs, I think we will see a race to the bottom in which there will be few winners.

One reason for this is the incredible ease with which content can now be created. A report from Hubspot found that just under half of marketers were using Gen AI to create content (although just 5% use it to write copy entirely from scratch). Hubspot also found that 75% of marketers said generative AI helps them create more content than they would without it.

Are business audiences clamouring for more content? I don’t think so. It’s not as if there is a shortage. Producing more might have short-lived benefits but it’s unlikely to be sustainable. Quality is another issue to bear in mind. Gen AI content can be out of date, biased, or just plain wrong and riddled with so-called hallucinations.

So if producers of content and thought leadership were to take a truly audience-centric approach to the adoption of Gen AI, what would this look like? I believe there are several important benefits of Gen AI tools to the audience that marketers should consider ahead of the efficiency gains.

1. Personalisation

One of the key benefits of Gen AI tools for marketers is the ability to create hyper-personalised content at scale, and use rich stores of customer data to target tiny segments with highly contextualised, relevant messages. This can also have huge benefits for the audience, by giving them access to content that genuinely meets their needs and is tailored to their specific circumstances.

An example from the world of consumer goods, quoted in a recent FT article, highlights the potential. Cadbury in India launched a recent campaign featuring Bollywood star Shah Rukh Khan, which encouraged viewers to shop locally. Using AI, thousands of local stores could customise the ad to feature the name of their own business using video footage of Rukh Khan.

In a B2B thought leadership context, it is easy to imagine how marketers could create highly tailored content outputs in a way that has not been previously possible. Imagine, for example, that your company produces a large, global, cross-sector thought leadership report each year that is an integral part of your brand messaging and narrative. If you do something like this, you’ll be used to getting dozens of requests from different regional teams for data cuts and tailored content, and probably to complaints that these big flagship reports are aimed at too broad an audience to be relevant.

Gen AI can solve this challenge by adding that layer of personalisation without manual localisation or customisation. Need a report for CFOs of consumer goods companies based in Brazil – one version in Portuguese and another in English? Or a video from your CEO customised for key accounts? No problem.

This does of course raise some ethical questions. When the Reuters Institute asked media leaders about which uses of AI carried the biggest reputational risks, by far the biggest (cited by over half of respondents) was content creation. This is not surprising, given the risk of disinformation and potential for content to be skewed by bias. Branded content producers will face similar ethical dilemmas – and great care and attention will need to be paid to ensuring that AI-generated content does not cause serious problems, either through disinformation or the accidental use of third-party IP.

2. Access

In addition to personalised content creation, AI tools could also help with ensuring that audiences are served content that is relevant to them. In the same way that Netflix or Amazon collects data about preferences and uses it to create recommendations, so B2B companies could apply a similar approach, provided they have the scale of data to analyse. Visitors to a company website could be served content that is tailored to their demographic profile (e.g. role, sector, company size or region) or to behavioural factors, such as stage of the buyer journey.

The way audiences search for and locate content will also be transformed. Rather than relying on traditional search which requires audiences to go through a long series of links to find a response to a search query, tools like ChatGPT provide a single answer to a question containing excerpts from multiple web pages. In time, this could make the standalone thought leadership report or series of articles obsolete. Rather than expecting the audience to piece together an answer to their question from multiple reports and articles, Gen AI aggregates it into one place, providing highly tailored responses to very specific queries. The days of flicking through multiple web links on Google, or browsing your Insights page in the hope of finding something relevant, will be long gone.

The days of flicking through multiple web links on Google, or browsing your Insights page in the hope of finding something relevant, will be long gone

3. Customisation

We know that audiences want to consume content in a variety of formats and lengths. Some of these preferences are down to learning styles: some individuals prefer visual content, while others would choose auditory, written word or kinesthetic (i.e. experiential). Audiences also want content at varying levels of depth. Some may be happy to read a long-form report, whereas others want more succinct answers to their questions.

Satisfying all these audience needs used to be a complex and expensive task, but Gen AI will in theory make it a lot easier. Long reports can easily be summarised into pithy articles, audio or video content can be converted into written word and search queries can synthesise points across multiple reports and formats. By providing variety and choice across formats and content lengths, we serve our audiences better, making them more loyal and likely to return to our content in future.

4. Timeliness

Audiences are hungry for insight that helps them to be more informed and make better business decisions. Providing them with this insight is usually a time-consuming endeavour for marketing teams because most forms of research are complex and manual. But what if you could hugely accelerate and automate research timings, getting insight in front of audiences on a much more frequent and timely basis?

One example of how this could change is the gathering of quantitative survey data. As producers of thought leadership know, this is an expensive and often lengthy process, but new techniques such as the application of synthetic data could enhance, accelerate or even replace some traditional survey methodologies. Synthetic data essentially simulates human responses to survey questions, using datasets that are already available. Although still in its infancy (and currently better suited to consumer research than B2B because datasets are a lot larger), there is a lot of excitement about its potential. One academic research paper, quoted in a recent column by Mark Ritson in Marketing Week, found that there was a 90% similarity between a perceptual map analysing car brands created using human respondents, and the same map created using GPT4. If this kind of accuracy is easily replicable, it could dramatically change market research – and by extension, survey-based thought leadership.

Imagine if, rather than commission surveys that take six weeks to field, you could get comparable data that was 90% accurate almost instantly? Time-to-market on thought leadership studies could be greatly compressed, which is good for marketing teams but also good for audiences, because they get access to insight so much more quickly.

Time-to-market on thought leadership studies could be greatly compressed, which is good for marketing teams but also good for audiences, because they get access to insight so much more quickly

An audience-first approach to AI is the only way ahead: audiences get timely, personalised, customised content that they can easily find, while producers save time and effort – and build a more loyal following. This is not to say that automation and productivity hacks are not valuable, but that there is so much more to gain by placing the audience at the centre of AI initiatives and considering AI innovations first and foremost through that lens.

This article was original posted on LinkedIn. Join the conversation here.