Take a B2B technology provider, which has evolved from a hardware vendor into an integrated systems provider, for example: The software is on the market, the website has been updated, and the sales team is talking about end-to-end solutions. Yet, ChatGPT, Perplexity, and other AI systems still position the company as a hardware vendor. When asked about integrated solutions, AI systems might even list the company’s competitors. This is a familiar communications challenge transported in the era of AI: The offering has changed, but the public perception of the provider has not.
A preliminary Generative Engine Optimization (GEO) check can identify where the AI systems’ perception diverges from the brand’s current profile. The next step is then to interpret the results, develop and implement appropriate communications strategies, and monitor how the brand’s image in AI system responses evolves.
This might sound like a major GEO undertaking, but in practice it can often start with adapting existing content and conducting targeted communications measures. Schwartz PR combines GEO analysis with more than 30 years of experience in technology communications to achieve exactly that. Our PR consulting and editorial teams develop tailored content across all channels, and position topics that are relevant to a company’s target audience as well as AI systems with the media through long-standing media relationships. The following four steps show how this translates into a coherent GEO strategy.
From positioning to prompts
A company’s visibility in AI responses can only be assessed within the context of a specific search scenario (searchintent): Does the brand appear in AI system responses to the questions (prompts) that potential customers ask in a specific research or decision-making scenario?
Prompt suggestions automatically generated by Large Language Models (LLMs) or GEO tools can provide a jumping-off point, but these prompts do not necessarily match the questions and phrasing that potential customers actually use when searching. Which is why Schwartz PR designs prompt sets for preliminary GEO checks based on a target audience’s actual search behavior, taking into account the brand’s overall context.
Prompts generally differ depending on a (potential) customer’s level of knowledge and their stage within the customer journey:
- Branded prompts show AI systems describe a company that is already well-known.
- Unbranded prompts test whether the brand is mentioned as part of an open-ended search for providers or solutions.
Both perspectives matter: A company may be accurately represented by AI when directly asked, but might still be missing from general comparisons between providers.
For our example company from above, some example prompts could be:
- Unbranded, understanding the problem: How can I catch signs of bearing damage in my pump?
- Unbranded, searching for suitable providers: I need to retrofit predictive maintenance capabilities, but machine data is to remain on-site. Which providers can work around that?
- Branded, evaluating the offering’s suitability: Can [company]’s software also analyze data from our existing sensors?
- Branded, preparing for an investment: How much does a five-machine pilot project with [company] cost? And how much would we pay for the software afterward?
Once an appropriate prompt set is designed, Schwartz PR uses it over an extended period of time to assess how frequently and prominently a brand appears in the AI system responses, which services AI associates with the brand, and which sources and competitors are cited more frequently.
Understanding what GEO KPIs mean for your brand’s positioning
A GEO tool shows how often a brand appears in AI system responses to a predefined prompt set, how high it ranks among all mentioned brands, and which services an AI system associates with it. Many tools also provide ready-made recommendations such as “Publish ten blog posts on predictive maintenance”, which may not fit the task at hand – for example, a GEO tool might recommend publishing content on competitor websites simply because AI systems frequently cite those as sources.
That is why Schwartz PR evaluates GEO metrics and recommendations in the context of a client’s actual communications goals. Our experienced consultants and editors examine which information is actually missing, which message needs to be refined, and through which channels it is best communicated.
Speaking of channels: When selecting which publications to place content in, a company should look beyond how often a publication is cited in AI system responses. For example, a specialist publication might block AI crawlers, but still be essential reading for this industry’s decision-makers. Therefore, we take a publication’s role as a potential cited source for AI systems as well as its reach, editorial relevance, and market position into account.
At the same time, we also assess a client’s own channels: Could a potential customer find all of the information needed here? Do a company’s website, blog, and press materials describe the company’s offering in consistent terminology? And is the relevant content publicly available and indexable?
For all these factors, the predefined prompt set serves as the benchmark: It shows whether a company appears in search results for questions relevant to its target audience, and whether AI systems associate it with its desired positioning.
Aligning communications activities with GEO
AI systems do not just consider a single source of information. They evaluate a brand within the overall context of all available sources. Therefore, content across various channels must paint a consistent “big picture” to effectively influence both a brand’s visibility to AI and positioning.
Essentially, content provides AI systems (and human readers) with indicators of practical experience, subject-matter expertise, industry recognition, and credibility in regard to a brand’s own statements. Google summarizes these “GEO virtues” as E-E-A-T: Experience, Expertise, Authoritativeness, and Trustworthiness. For Google, this means that the more clearly various sources demonstrate that a company has the sought-for competence, the more likely AI systems are to correctly attribute it to the company in their responses.
With this in mind, Schwartz PR tailors content and channel selection to achieve the intended overall effect. Our EPOS (Earned, Paid, Owned, Shared) channel model classifies potential communication measures as editorial coverage (Earned), paid placements (Paid), owned channels (Owned), and public engagement on social media and in professional communities (Shared).
For our example company, these could work together as follows:
- Owned: A comprehensive landing page explains how hardware, software, and services interact, how its solution is integrated, and what the solution can do in practice. Product pages and other content use the same terminology and key messages to ensure consistent messaging.
- Earned: A user case or expert interview in a leading industry publication makes the company’s solution understandable through a concrete example, providing potential customers with technical context, which Asystems can cite as a trustworthy source, even beyond the company’s own website.
- Paid: Through a media partnership, the above-mentioned user case can also be presented to an additional audience of industry professionals. If the content remains publicly accessible and indexable, this can also serve as a cited source for AI system responses.
- Shared: The company answers practical questions on LinkedIn and in relevant industry forums, also providing links to more in-depth information. If there are follow-up questions, these can highlight areas, where further explanation is required.
If, for example, an initial GEO check has surfaced that a company’s software expertise is barely visible on its own website (Owned) and in cited technical articles (Earned), Schwartz PR expands the existing product solution webpages with appropriate content and a tailored user case specifically to address this issue.
Tracking GEO results and letting them inform the overall strategy
Once individual GEO measures are implemented, Schwartz PR then uses the prompt set defined at the outset to assess how a brand’s positioning within AI system responses has evolved.
GEO and media monitoring are closely intertwined for this process: Which topics and messages are media editorial teams picking up on? Are the articles published by them used as cited sources by AI, and are their statements reflected in AI system responses? On this basis, we can assess a company’s visibility to AI systems in the context of media coverage.
Whether communication measures are also reaching the right potential customers can also be assessed using website data and feedback from sales teams: Is traffic to the relevant product solution webpages increasing? Do inquiries refer more frequently to a specific software offering? Do potential customers cite AI research as their starting point, or do they already enter the conversation with a better understanding of the company’s solution?
These assessments are then fed back into further planning processes. Schwartz PR works with clients to determine which communication measures have proven effective, where content needs to be refined, and which topic should be focused on next, ensuring that GEO remains aligned with the rest of a client’s communication: What should the company be known for, who does it want to reach, and what does it want the audience to know and understand about its offering?
GEO as part of existing communications
In practice, GEO expands existing content and public relations work by adding an additional perspective, enabling targeted prioritization of planned website updates, expert articles, case studies, media contacts, and social media content. That is why Schwartz PR combines a GEO ranking check with optimization of existing content, GEO-first content, media relations and continuous reporting – with the goal of GEO as the same as that of any communication: Ensuring that a company appears in the search results for questions that its potential customers are actually asking, and in sources those potential customers trust.