Large Language Models (LLMs) remain a hot topic in SEO – especially with the popularity of OpenAI’s ChatGPT and other generative AI tools with user-friendly chat interfaces.
It’s so easy to get excited about the potential of generative AI and what it means for an SEO strategy. But managing client expectations remains critical.
This overarching guide serves to help SEOs educate clients on the potential functionality of LLMs as it applies to SEO from the client perspective. This pragmatic approach will help you set and maintain realistic expectations throughout your SEO projects.
Preliminary steps: Ask the client the right questions
Some clients come to the table operating from info collected in word-of-mouth interactions, social media posts, and headlines rather than a detailed understanding of what AI is and how it works.
As the expert in the room, the SEO maintains the responsibility of naming both the benefits and drawbacks of using AI tools for organic search initiatives.
Asking the right questions before jumping in too far clarifies objectives and serves as a preliminary risk assessment for all parties:
Assess client familiarity
Gauge the client’s knowledge of LLMs (as it pertains to SEO) to ensure effective communication.
If the client brings up the topic, ask leading questions about what inspired the ask – whether it was an intensive course, first-hand experience, or a stakeholder brainchild. The conversation allows SEO to steady the course.
Understand client desires around ROI
Determine what the client hopes to achieve in terms of return on investment (ROI) by using these tools.
Learn about the potential investment in enterprise or developer-owned versions of the tools that can offer more in terms of privacy and/or quality output.
Clarify client objectives
With more context, help the client shape their specific goals for using generative AI in their SEO strategy if unclear.
Empirical testing: Measure feasibility, quality and time commitment
Conduct empirical tests to assess the effectiveness of generative AI for a given task.
To get started, outline the steps involved in conducting empirical tests, such as:
- Setting up test environments (which LLMs are being used).
- Creating and experimenting with prompts.
- Evaluating the quality of generated responses.
Evaluate the feasibility, quality, and time efficiency of LLMs compared to manual methods.
Ensure that these metrics are directly tied back to any initial objectives set, providing a complete feedback loop in your project management.
Offer a formal point of view (POV) before initiating work
A well-crafted POV document can help set clear expectations for clients.
A typical POV on using LLMs in SEO could include the following sections:
- Introduction: Briefly explain the purpose of the POV and its importance in answering the client challenge at hand. Name the problem the client aims to solve and frame the usage of AI as a possible solution.
- Competitor/landscape research: Look to competitors to determine how they are integrating AI technologies into their SEO and digital marketing processes. Whether the use case is backend or user-facing, a competitor who is already using the technology offers a glimpse into what looks to be pure hype versus what produces results.
- Capabilities: Outline the specific features and benefits of generative Ai based on the client’s ask. Avoid oversharing jargon-ridden details unless necessary. It may be useful to include screenshots of outputs from the tool.
- Limitations and risks: Dispel the notion that LLMs are a cure-all solution by explaining their limitations. Note that LLMs are evolving, and today’s hurdles might not apply soon. Conversely, today’s capabilities could be scaled back due to legal or ethical implications. Discuss potential risks associated with using generative AI, such as privacy concerns, data security, or the possibility of generating inappropriate, inaccurate or biased content.
- ROI: Provide an estimate of the potential ROI based on the project’s objectives and scope. Consider naming the effort and impact of the work compared to doing it manually.
Creating a comprehensive POV document establishes a solid foundation for managing the client’s perspective and enables all parties to be on the same page before moving forward with the project.
Establish objectives and goals for generative AI and LLM usage
It’s no secret that effective marketing initiatives, SEO and otherwise, are built around clear objectives and goals.
Based on the client’s initial questions and the response to a POV, formulate Specific, Measurable, Achievable, Relevant, and Time-bound (SMART) SEO goals for your project to ensure the usage of LLMs is for a specific purpose.
- Define key performance indicators (KPIs): Identify metrics that will be used to evaluate success. Both SEO metrics and product output metrics might be considered based on client goals.
- Incorporate objectives and KPIs into an SEO roadmap: Develop a roadmap that outlines how you will achieve your goals.
Incorporate goals into an SEO roadmap
Using the findings from researching and testing the LLM tools, adding sensible deliverables to your SEO roadmap allows you to make necessary adjustments that align with your client’s overall goals and strategies.
Remember that setting and managing expectations is an ongoing process, so be prepared to adapt and adjust your strategy and roadmap as needed due to technological advancements and changes in the legal landscape.
With proper planning and communication, you can help your clients harness the power of LLMs while minimizing potential risks and maintaining a sustainable and ethical approach to SEO.
Look to the future in answering client questions
When addressing client questions about AI and LLMs, consider the future implications of implementing the tools in an SEO strategy.
Stay informed about industry trends and updates to make well-informed decisions about which tactics to employ.
For instance, while ChatGPT and similar tools can quickly generate FAQ and how-to blocks with associated schema, Google recently announced that this feature is being phased out for most sites. This emphasizes the importance of recognizing that the longevity of certain tactics may be uncertain.
In other cases, it is a matter of considering ethical concerns for a suggested SEO strategy using something like ChatGPT.
For example, if a client intends to automate SEO content creation for numerous pages using prompts like “Write a 500–700-word search engine optimized article about Carpet Cleaning in XYZ City” without human supervision, imagine the potential impact on user experience and search engine reactions to similar content across multiple websites.
Pose questions such as:
- What happens if everybody does this?
- How would consumers react to seeing ten similar generative-text-created blogs on 10 similar websites?
- How might Google modify its system if/when this occurs?
This line of thinking can help prevent wasted resources and ensure a more sustainable approach to SEO.
By following the steps outlined in this article, and maintaining open lines of communication with your client, you can foster stronger relationships and ensure the long-term success of your SEO projects.
The post SEO, generative AI and LLMs: Managing client expectations appeared first on Search Engine Land.