The current state of using AI for SEO in 2023
I’ve got some interesting data to share with you on our current status as an industry when using AI for SEO.
Over 860 digital marketers signed up to share their opinions and experiences on using AI for SEO. All respondents are current users of the Screaming Frog SEO Spider.
10% have never used AI before, and 90% have used AI.
How can I gain client trust and convince them of AI’s ROI?
If clients don’t know when you are using AI, they’ll feel cheated and suspicious of your future service. Transparency is telling your client how you use AI (such as for keyword research) and the impact on results.
If you mention in passing:
“I’ve used ChatGPT-4 to help with the attached keyword research.”
Using AI for keyword research could indicate to clients that you’re not the expert you claim to be. Most generative AI have a freemium model, so they assume anyone can use it for keyword research.
However, if you’re specific and say:
“I’ve used ChatGPT-4 to help generate a more in-depth seed list of terms, alongside brainstorming my own Google & YouTube autosuggest, scraping PAAs, and analyzing competitor pages.”
Clients will likely have more trust in your ability since you’ve demonstrated you’re using AI as a wider mixed tool approach for a seed list of keywords.
Particularly pertinent for agency-side SEOs who charge daily rates, you must be accountable for your work hours. Accountability showcases how using AI in your workflow increases ROI and avoids client escalation.
If AI saves an hour of work daily, that accumulates to 5 hours a week and roughly 20 hours a month. Clients may question where that time is being spent. However, this doesn’t change that clients are paying for your expertise, and this should be emphasized.
SEO is a results-driven industry, and if it’s necessary, then billing strategies should incorporate results in a tailored performance contract — this way, using AI or not, all parties agree on the common goals to reach.
Alternatively, if this isn’t how your agency model works, here is how you should handle this:
Time tracking: Ensure all employees are using tools like Toggl to understand how much time AI saves them per task and feed that into estimates of the pre-defined scope of work OR use it to inform flexible billing.
Pre-define the scope of work: Clearly outline the scope of work and itemize each activity at a daily rate. Be honest on tasks streamlined through AI.
Flexible planning: If using AI varies significantly, you could opt for a flexible planning approach based on employee time tracking pro rata.
Employees can spend extra time on other tasks, further highlighting more ROI for using AI.
I was shocked that 24% of SEOs using AI admitted not understanding how LLMs work. We’ve repeatedly seen SEOs gravitate to the latest SEO tool without thoroughly interrogating its use.
If it looks nice and saves me time, it must be good, right?
Taking courses in LLMs will help you spot the main limitations of AI, and clients are within their right to question your experience with AI, as they will be the ones to suffer.
I’m only touching the surface of these limitations, but here are some examples:
Data privacy concerns: Do not share sensitive data with ChatGPT, Bard, or equivalent. In SEO, this relates to sensitive traffic and revenue data, client staging content, or budgets. Anything you input into LLMs may eventually inform the broader LLM model. Hence, you should adopt an intelligent strategy for what you share with LLMs. But don’t let that deter you from using advanced data analysis and Code Interpreter on publicly available data.
Inaccurate outputs will happen: Generative AI is only as good as the prompts and subsequent iterations you feed it. Learn how to iterate for different tasks, but know that there will be occasions where its final output is not focused enough on brand values or is of high quality to use.
Social biases may exist: As a brand, you have a responsibility to uphold core values, especially equality. Generative AI might have biases if the training data was predominately skewed to a proportion of individuals. However, by acknowledging pre-existing bias, you can mitigate and produce better outputs that inform your SEO strategy.
I’d suggest taking some generative AI and LLM courses that help you demonstrate your education to clients. Some free courses at Google include:
Boost Your Productivity with AI
Introduction to Generative AI
Introduction to Large Language Models (LLMs)
Share your credentials with clients, and they’ll be more convinced that your handling of AI will lead to a higher ROI.
What makes an ideal AI compliance, privacy, and ethical framework?
I was delighted to learn that 62% of respondents already have an ethical framework. Before I get into more details, I want to define what I mean by ‘Compliance, Privacy, and Ethical Framework.’
An ethical framework ensures clients and agencies align on the approach, which helps you get buy-in. You must share this framework with clients, ideally at the onboarding stage.
A convincing AI framework should include:
1. Your approach
Split your approach between onsite SEO and offsite SEO if relevant. Give a holistic overview of how you’d typically approach using AI to support these pillars of SEO.
At this stage, listing every task you use AI for might not be necessary, but provide details on how you’d approach it. For example, your approach may be across onsite SEO to help with inspiration (perhaps across seed keywords, metadata, and more). Or, you may use AI for general data analysis, organization, and categorization.
Telling clients you use AI for ideation indicates you’re not completely reliant on AI for onsite analysis.
If you prefer to give a task-by-task breakdown, that’s your prerogative. But with SEOs utilizing new innovative AI techniques by the day, this might save you some effort to update it regularly.
Additionally, you should include a complete list of different generative AI tools you lean on. For example, this may be:
This gives clients more transparency into any tools they might have issues with or those that align with their framework.
They may even have access to AI tools you don’t, which will open up a dialogue. For example, perhaps they have ChatGPT Enterprise access, which may limit the data privacy concerns we mentioned earlier. Then, you can adjust your approach accordingly.
Also, include details on how AI impacts billing, whether through a flexible billing strategy or pre-defined project allocations based on experience using AI for similar clients.
2. Privacy assurances
Privacy assurances are a list of data you’ll never input into AI tools and are based on your knowledge of your AI tools.
For example, if you don’t have a ChatGPT Enterprise account, data privacy will be a more significant concern, and you’ll need to provide assurances on what data you won’t share. It could include:
Sensitive traffic & revenue data from GA, Adobe Analytics, etc.
Internal brand guidelines or marketing information
Target audiences and segments
3. Quality assurances
Be precise with your wording here, you might not need to give a comprehensive list of assurances on individual tasks, but instead of mentioning you “may use AI for content creation”, specify it’s more about helping shape content based on the tone of voice, spelling & grammar, content briefs and so on — something covered extensively within this “Make AI Your Writing Sidekick” article.
Ensure you emphasize your education with any agency-wide AI accreditations if applicable. If not, highlight that all work will undergo multiple iterations and be reviewed by experienced SEOs for quality control.
Sharing that you’re committed to staying up to date with new techniques and Google’s best practices for AI would be wise.
4. Client feedback
Leave some room for open-ended feedback or a checklist-based questionnaire on a client’s experience.
Here are some questions that are worth considering:
Can we use our highlighted approach for your SEO strategy?
Do you have any internal AI tools at your disposal?
Can you share AI tools during our engagement?
Are there any AI tools you’d like us to avoid entirely?
Has any AI-based work gone live that we should be aware of?
Do you have any further thoughts on using AI for SEO?
Asking the above ensures you comply with their internal strategy and identify any potential concerns before they escalate.
Embracing Ethical AI in SEO: A Path to Transparency and Trust
Using AI has undeniable benefits, primarily increasing productivity and efficiency while enhancing your expertise across SEO.
However, proceed with caution. It’s imperative to consider the wider ethical implications of using AI, such as data privacy, quality, and social biases. To further pave the way for ethical AI implementation, it’s advisable to add a framework that details your approach to establish stronger client trust.
To help digest the information given within this article, I’ve provided an infographic to share with colleagues and SEOs: