AI is increasingly used by B2B buyers. This article explores ways to create content to influence their responses and make you appear in search results.
An Observation: AI is Increasingly Used to Research B2B Companies
It’s a fact: more and more internet users, and especially B2B buyers, are looking for information about your company and your services online… but above all, since the advent of ChatGPT, on « AI search engines » (The Hidden B2B Buyer Journey: Deciphering Decision-Makers’ New Habits). Even if B2B purchasing is more complex and the sales process doesn’t always end on an e-commerce site, your future clients still form an opinion about you through the answers they get from ChatGPT, Gemini, Claude, or even Mistral, as well as from your own site. Both are important and both are part of the buying process.
To be honest, it remains difficult today to measure a presence and predict what ChatGPT will say about you. Tools exist for this, but their usefulness is questionable. Nevertheless, concern for these LLMs must be taken into account starting today, because all figures show their growing use in opinion formation, and your company will not escape it, even if you are in a small market.
But how do AIs form an opinion about your company? And can we really try to influence AIs for this?
To answer this, I will rely on an article by Jason Barnard, an American SEO specialist, who writes a series on these questions, and more specifically on his latest article: How AI forms opinions about your brand.
The 3 Dimensions of Information that Feed Generative AIs
Jason Barnard defines three types of information used to describe content that can feed AIs and make you appear in their responses. Don’t forget, at this stage, that AIs don’t just feed on one data source to form their answers, but on all information across all digital media they find about your company: your site, your social networks, but also others’ social networks and other sites (Reddit: the new Eldorado of digital marketing?). Therefore, there is a part of your information sources that you do not control and that does not stem from content production work, but directly from the quality of services you provide to your clients.
1 – Identity: Who are you?
Before answering an internet user, an AI must clearly know who you are, what you do, and what types of companies you address?
Most companies already answer this question through their « About Us » page (see my article on this subject), but also through all product and service pages, whether you have an e-commerce site or not.
2 – Credibility: What proves that you do what you do well?
Here we enter a predominant domain. AIs will only cite you if they are convinced that you are credible in what you do.
Even if this partly depends on third parties, nothing prevents you, quite the contrary, from asserting this credibility. This involves well-known artifacts that many companies already implement.
- User cases
- Testimonials
- Qualifications/certifications
Read also: GEO: In the age of AI agents, take care of your identity page
3 – Recommendation Capability
Does the AI engine have the necessary content to recommend you to users who match your audience? This is where things get tricky for most companies, as they are unaware that this capability comes from content they mostly do not exploit at all.
This obviously comes from the content you already publish.
- your thematic content;
- your marketing actions;
- the authoritative content you produce or have produced.
But the recommendation capability often hides where we don’t look for it: in the content naturally generated by the company’s operations and by offline activities.
In other words, companies often already have much of the information AIs need to understand, trust, and recommend them. The real challenge is less about creating ever more content than about identifying, structuring, and making visible the knowledge already produced daily by the organization.
4 Data Sources to Fuel Identity, Credibility, and Recommendation Capability
Here, Jason Barnard revisits 5 information flows that you can control and that help feed the 3 dimensions of information examined by AIs.
- Products and services
- brand
- expert content
- operational data
- offline
1 – Products and Services
Most companies already do this, but you can still deepen your content: who it’s for, what problem it solves, what it doesn’t do, what’s the difference with other similar products or services.
2 – Authority/Expertise
These are often the contents you already publish: articles, videos, data, guides, studies, expert speeches.
The Brand Voice
The voice is probably the element that companies master the least.
Brand narrative corresponds to what you say. Voice, on the other hand, corresponds to how you say it.
A team can write the narrative once, but the voice is expressed everywhere: in sales exchanges, customer support responses, social media posts, sales presentations, and all content produced by the company.
When this voice deviates, AI perceives the same brand as if it were five different brands. The result: it loses confidence in each of them.
This is why it is essential to standardize your brand voice and maintain its consistency across all your touchpoints.
3 – Business Operations
The resource that is never used, but is the richest.
These are daily data.
The idea is that companies possess an immense amount of unique information generated by their daily activity:
- customer exchanges;
- support feedback;
- specifications;
- project reports;
- audits;
- internal studies;
- usage data;
- real FAQs;
- business procedures;
- experience feedback;
- production or sales data.
This information is often proprietary (it exists nowhere else) and therefore constitutes a very strong signal for AI engines, as it demonstrates real experience and concrete expertise.
Obviously, the difficulty lies in collecting this information. In large companies, it is often complicated because the fear of communicating confidential information is strong, as is content control. This makes internal information gathering operations complex and lengthy.
One must also rely on the willingness of those who create this information. Often, they do not see its benefit and consider it a waste of time.
A strong directive from management must be initiated to enable this collection work.
I have implemented a process that allows this to be done quickly by organizing rapid interviews and rewriting this content using AI (Content writer offer).
Jason Barnard suggests implementing a tool for centralizing this data. This is a separate project, but one that can very well be carried out by internal teams using no-code tools like Timetonic.
Timetonic: a no-code tool for all needs
Timetonic is a no-code tool for structuring and managing databases. It is very easy to learn and does not require advanced or specialized technical knowledge. Thanks to Timetonic, you can build a content database that can then be used across all your digital channels.
4 – Bringing Offline Online
Finally, there remains one last data source little or not exploited by companies. These are the data created during offline meetings: breakfasts, after-work events, trade shows, conferences.
Many conversations, many things can be transformed into content and re-used to create online content.
This works both ways, by the way. Online can also very well feed offline. The important thing is to connect all of this to strengthen the credibility of all this content with LLMs.
Conclusion
Jason Barnard’s article is much more comprehensive than mine and is surely much more adapted to all types of businesses. I wanted to focus here on communication for B2B SMEs-ETIs. We find exactly the same principles there. The only difference lies in the organizational capacity of these companies.
One thing is certain and must be well understood: if this communication strategy is AI-oriented, it is equally beneficial for humans. In other words, it’s not about telling yourself that you’re doing it only to feed ChatGPT or Claude, but also to feed your prospects and clients, as well as your SEO. This forms a whole encompassing a set of users: humans and machines alike. In the end, however, it is indeed your future and current clients whom you serve.
Such a strategy can only be long-term, but it will be much more rewarding than paid campaigns that stop as soon as you turn off the tap. In any case, the two methods are complementary.
Do you want to implement such a strategy? Contact me and I will help you build your future brand.
Laisser un commentaire