1. Who will give the answers? Generative AI might<\/h2>\n
Generative AI has become a significant topic for brands worldwide. <\/p>\n
The growing number of users researching AI-powered platforms gives us an idea about just how big the issue may become.<\/p>\n Even non-marketers use chatbots linked to platforms such as OpenAI to ask everyday questions. This is made much easier by chatbot platforms like Botpress or Messagebird.<\/p>\n The search landscape is changing heavily. As a result, online reputation management (ORM) will never be the same.<\/p>\n In the past months, numerous tests showed critical shortcomings, such as: <\/p>\n A sustainable strategy to protect your brand from generative AI issues is to control the input. Language models are very likely to use various sources on the web to feed their algorithms to get text samples and understand the facts.<\/p>\n A brand (business or personal) that triggers a knowledge panel on Google is very likely to be in a good place when it comes to the semantic part of ORM. <\/p>\n Generative AI will use more of its internal knowledge management systems to produce more accurate answers, considering real and understood facts.<\/p>\n To proactively control the narrative, it’s critical to work on connecting your brand to appropriate semantic fields to make it easier for language models to comprehend:<\/p>\n In practical terms, this might mean tying the brand’s context to descriptive terminology around the skill level, price point or specific targeting:<\/p>\n As so often, input = output<\/strong>. The more you control the narrative on your website and other sites about your brand, the higher the chances that language models pick up the right information and reproduce it in the right contexts in generative answers.<\/p>\n Semantics aren’t always clear: It often takes multiple semantic entities to be presented in the right context to allow NLP engines to actually comprehend what the text is about.<\/p>\n The free Kalicube Copywriting Analyzer Tool allows everybody to ask Google’s NLP engine through their API about its understanding of a sample text.<\/p>\n Look at these examples and Google’s understanding in parentheses:<\/p>\n These examples illustrate how hard it can be for neurolinguistic processing engines to grasp what a text is about.<\/p>\n Context is king when it comes to getting an NLP engine to comprehend what your page is about. Make sure <\/strong>your content is tied to enough context to be unambiguous while paying attention to different spellings, synonyms, uppercase and lowercase versions, etc.<\/p>\n Just because there is text on a page that a human can fully understand doesn’t mean a machine can take away the same information.<\/p>\n Online reputation management needs to focus on what gets the attention online. Very often, user-generated content<\/a> (UGC) accounts for a major share of that. Users generate content in different formats, such as YouTube videos, podcasts, images and posts.<\/p>\n Given Google’s partnership with Reddit, the widespread community network has become a critical priority for ORM.<\/p>\n Many branded and product-related searches will trigger one or more Reddit threads<\/a>. In addition, Google’s Gemini could leverage Reddit’s content as training data for its language models.<\/p>\n Actively monitor Reddit and maintain a few official users representing your brand. Look into creating new threads and start participating in existing discussions around your brand and contextually related topics.<\/p>\n To control the narrative (and the training data for a potential SGE roll-out), maintain a positive and helpful attitude and try to lead users with negative experiences off Reddit and into your email support desk where you provide real help, compensation or other benefits to minimize the damage.<\/p>\n Dig deeper: How to repair your Google search results and reclaim your online reputation<\/a><\/em><\/strong><\/p>\n The more popular brand, the more mentions there are on the web. <\/p>\n Tools like Google Alerts can enable an easy and free entry point for monitoring mentions of a brand (add different spellings) and its products or services. <\/p>\n However, Google Alerts is limited to indexed pages only. It will alert you by email about what is coming up in Google Search, while leaving you in the dark about most social media posts, threads and comments sections.<\/p>\n Effective brand mentions monitoring requires using (usually paid) third-party tools that search through a variety of platforms and consolidate all mentions in a user-friendly report, ideally with an indication of the sentiment of the different mentions. <\/p>\n But here is why monitoring mentions is even more complex than that:<\/p>\n How should a brand possibly know what kind of information is being presented through generative AI on Google and all the other major platforms out there?<\/p>\n Brands can hope Google might provide some kind of tracking through Google Search Console, but at the moment, it only seems that Google SGE heavily impacts many brand and product searches<\/a>.<\/p>\n If a brand wants to know how it is being portrayed across the different platforms, the simplest way is to query them manually (or set up an integration using third-party tools or Google Sheets plugins, etc.) to help with that. It’s not perfect, but it might give you hints on what to do next.<\/p>\n There is no excuse for not configuring a series of the most important Google Alerts to monitor your brand name, your main products’ names and their most essential spelling variations.<\/p>\n If your brand is frequently mentioned on social media platforms, you should sign up for a paid tool to cover content that Google might not index.<\/p>\n Besides, it wouldn’t hurt to assign a team member to frequently query the most important AI tools with defined transactional queries about your brand and products to understand the topics and applications associated with the brand. Then, react accordingly to their output by finding the likely source of unwanted information and aim to balance it out.<\/p>\n Get the daily newsletter search marketers rely on.<\/p>\n \t\t\t\t\t\t\t![\"AI](\"https:\/\/searchengineland.com\/wp-content\/seloads\/2024\/05\/AI-platforms-Trend-data.png.webp\")
<\/figure>\n\n
\n
\n
2. Ambiguous semantics<\/h2>\n
\n
3. User-generated content and Reddit<\/h2>\n
4. Lack of proper monitoring<\/h2>\n