The hype around AI is palpable.
Every day, it seems like AI makes a new leap forward—the release of ChatGPT-o1 may well already be old news—and with each leap, it seems like there’s a new prediction about AI tools replacing writers, strategists, you name it.
We get it—the possibilities are exciting. Recently, we published an article talking about how, despite all the breathless predictions, AI isn’t about to change everything.
However, in this article, we’re going to elaborate on that… by talking about all the ways AI has changed our content marketing process at TELL ME MORE®.
But along the way, we’ve learned a valuable lesson: AI is a powerful tool, but it’s not a magic bullet.
AI can help our writers and strategists with ideation and brainstorming, providing a springboard for creativity and helping outline the structure of our stories. Yet, the heart of our content—deep research, insightful interviews with real people, conveying contextual emotion, and the nuanced decisions that shape our narratives—remains firmly in human hands.
It’s these elements that breathe life into our content, ensuring that our narratives resonate deeply and our brand voices stay authentic and engaging.
In this article, we’ll break down the TMM content marketing process. You’ll see how we use artificial intelligence strategically—and why the human touch is still essential for producing content that truly resonates with readers.
Looking to strengthen your SEO and content marketing strategy? Let’s connect—submit a request or schedule a free consultation to see how we can help you grow your audience and drive conversions.
AI Is a Differentiator—But That’s Not Always a Good Thing
We believe AI has the potential to make the best marketers and writers even better. But in the wrong hands? It’s a recipe for generic, uninspired content.
Why? Well, as an internet user you’ve almost certainly already experienced for yourself the disappointment and frustration of being confronted with content that was either heavily or exclusively produced via AI. You can tell when you come across it. It’s vague, and doesn’t sound credible. It rarely (if ever) cites authoritative human or documentary sources. Its tone is just… off. And above all, it gives you the impression that it was just created to rank in search engines, exploiting your clicks rather than serving your search intent.
Google has also recognized the need to combat the proliferation of low-value AI generated content. In March 2024, the company released one of its biggest core algorithm updates, alongside various spam updates, to address the increased volume of low quality, unoriginal content appearing in search results. The swift penalties led to the de-indexing of many websites that relied solely on unoriginal AI-generated content as a quick means to climb the search rankings.
All told, in the hands of substandard content creators, AI’s greatest contribution has been in shining a light on the weaknesses in their process and presentation.
Contrast that with Gartner’s recent prediction that, by 2026, “80% of advanced creative roles will be tasked with harnessing [generative AI] to achieve differentiated results”—and that this will actually cause marketing teams to spend more on creative human talent in order to stay competitive.
That’s because high quality content creators connect with audiences by speaking to the fundamental principles of human psychology and needs in their writing. It takes a human touch to do that. And it takes hard work. AI can play a role—but it’s a supporting one, rather than the lead.
With that said, let’s walk through the article production process we follow at TELL ME MORE®, and take a look at exactly what is involved at every stage.
Step 1: Ideating Your Topics and Performing Keyword Research
Content production starts with content ideation, the process of choosing topics to write about. This varies based on the client because each one has unique goals and target audiences—but, generally speaking, two key ingredients are considered in this process.
- Client’s Goals: Understanding the client’s goals is crucial. Whether it’s increasing brand awareness, generating leads, or driving sales, the content should align with these objectives. Content ideation revolves around creating topics that resonate with the client’s aspirations and support their overarching marketing strategy.
- Search Intent of the Client’s Target Audience: Identifying the search intent of the client’s target audience is equally important. Search intent refers to the purpose behind a user’s search query. By understanding what users are looking for, content creators can craft content that directly addresses their needs and provides valuable information. This approach increases the likelihood of the content ranking well in search engine results pages (SERPs) and attracting the right audience.
Understanding Client Goals
For each client we work with, we build a keyword position tracking report. This report identifies the specific keywords that the client wants to rank for, which are typically aligned with their core products or services or a specific target audience.
Keyword position tracking reports include metrics such as monthly Google search volumes, keyword difficulty, and result positions, which help us identify both high-traffic keywords and underperforming ones. The latter is especially important, as it highlights opportunities to break into the all-important first search engine results page (or, ideally, landing in one of the top three search results).
Data varies, but the latest some recent research from Backlinko has shown that the first Google result boasts a click-through rate of 28.7% and more than 54% of all clicks go to pages that appear in the top three results ; by contrast, only 0.63% of all Google users clicked on results appearing on the second page.
In the example below, we created a report for a natural foods brand that captured keywords related to their core products—pork, ham, bacon, kielbasa, and so on. As you can see, there are over 350 entries on this list. That’s a lot of potential article topics. Each entry also includes search volumes and result positions, which help us focus our attention on both high-traffic and underperforming keywords.
So, evaluating a client’s position tracking report is one place we can start the process of ideating article topics that will help our clients meet their goals.
Understanding Search Intent
But client goals are only one half of the equation. We also need to think about what the target audience has in mind when they open their laptop or pull out their smartphone or tablet. What are they hoping to get when they type in a keyword? What is their intent?
Search intent is generally broken into four categories.
- Informational: Say you’ve recently heard about a popular cookware brand called Le Creuset, and you’re curious to learn more. You might search, “Why is Le Creuset so expensive?”. That’s an informational search.
- Navigational: If you know exactly where you’re trying to go and are only using a search engine to help you reach that specific page, you’re performing a navigational search—for example, if you’re trying to reach the Le Creuset website and search for “Le Creuset”.
- Commercial: If you’re thinking about purchasing a Le Creuset product but need more insight in order to make a decision, you might search for “Le Creuset reviews”. That’s a commercial search—it’s a bit like an informational search, except that you’re specifically looking for information that’s directly related to making a purchase.
- Transactional: If you’ve finally made up your mind to buy that Le Creuset product you’ve had your eye on since the beginning of this bulleted list and are trying to reach a page that will help you do that, you might make a transactional search—like “buy Le Creuset online”.
Search intent has an important bearing on both the type of content we create, and on the keywords we target when creating that content—in other words, it determines the objective of a new piece of content. But how do we learn what our target audience’s intent is? And how do we learn what people are actually searching for?
Ideating Your Topics (with the Help of AI)
Topic ideation is one area where AI tools can make the process more efficient. Let’s illustrate by returning to the example of the natural foods brand, and imagine we wanted to create content that would help our client rank better for keywords involving bacon.
AI’s Role in Ideation
We can head over to ChatGPT (or your preferred AI chatbot), give it a prompt with some details about our content goals, the client’s brand, and so on, and see what it gives back. It’s important to note that the more context you can give ChatGPT about your article’s target audience, objectives, SEO keywords, etc., the better the output.
In the example below, you can see we got several interesting ideas; we could quite possibly be looking at next month’s content plan—a series of articles all about bacon.
AI’s Limitations in Ideation
Of course, it’s not that easy. Simply taking AI’s word here isn’t going to help you create great content that will deliver results for your clients.
Why not? For one thing, ChatGPT is extremely adept at telling you what it thinks you want to hear. This means it’s prone to giving apparently logical responses that may not actually correspond to reality—the so-called “hallucinations” to which all AI models are susceptible. If you don’t manually fact-check its content, you’re likely to serve inaccurate information to your readers.
More importantly for the context of this process, ChatGPT can’t completely validate the search volume for the topics it presents you with (at least, it can’t do that yet). These articles might sound good, but is anyone actually searching for them? If it turns out there’s very little search volume, or that the topic doesn’t really fit into your target audience’s interest or intent, then it’s simply not worth your time. You can prompt ChatGPT to provide some suggested keywords—but you still need to go out there and determine whether those suggestions are valid.
The bottom line is that you still always have to go out and make sure your ideas match up with reality.
AI’s Value in Ideation
So, the value AI brings here is that it can help you quickly ideate many topics related to a core concept or keyword that your client wants to prioritize. This can save you valuable time trying to come up with—and investigate—unique concepts yourself. Even if you don’t like any of the responses the AI gives you, just going through this exercise might stir your own creative juices and help you come up with ideas faster than you would have otherwise.
But the bottom line is that you still always have to go out and make sure your ideas match up with reality.
Validating Your Concepts
Possibly the simplest way to validate your concepts is to take those ideas, hop on Google, and start searching.
Using your AI-generated topics and keywords as seeds, check the “People also ask” section to find common keywords that people are actually searching for—as in the example below. This can help you quickly branch out from your initial concept, gathering ideas to include in your content, picking up inspiration for article titles, or even identifying entirely new topics to explore.
This is a great approach, because a) it’s free, and b) you know people really are searching for these keywords. But you can go deeper than that, using dedicated SEO resources like Ahrefs or Semrush (we use both tools at TELL ME MORE® but prefer Semrush).
These platforms offer versatile tools that marketing agencies can use to dive deep into keyword research and come up with highly targeted content strategies.
We use Semrush to explore a variety of data points, including:
- Keyword volumes – how many people are searching for a term
- Keyword difficulty – how hard it is to rank for a term
- Keyword clusters – closely related keywords that, when used together, can make your content more efficient
- Keyword intent – the search intent most commonly associated with a keyword
- Keyword gaps – keywords that industry competitors are ranking for, but we aren’t
This allows us to put hard numbers to our abstract ideas and make confident decisions about the topics we want to write about.
Getting to the point where we can make confident decisions is what it’s all about. And even though a human ultimately needs to make those discerning decisions, technological tools—including AI—can help us get there more quickly.
Once again, any opportunity to save time helps us focus more on strategy—giving us an advantage over competitors, especially those who are either not using AI at all or who are using it poorly.
At the end of the day, AI requires a human being to actually make a decision. What’s the intent of my client? What are we trying to create? Decisions rely on subjective value judgments that vary from human to human. AI can provide knowledge, but human beings must provide answers.
Step 2: Organizing Your Content Plan
As we identify topics for each client, we integrate them into that client’s content plan, which we organize using Airtable.
Google Sheets is another viable option, but the advantage of using Airtable is that it’s extremely easy to create filtered views for specific clients, writers, article statuses, or any other important criteria. This is especially handy when working with a large team of writers; for writers, it makes it much easier to keep tabs on their current and upcoming workload, and for editors it makes it much easier to assign out work.
We typically plan content about two months ahead of scheduled publish dates. In terms of the workflow for each individual article, it’s broken out into several deadlines, with anywhere from two to six weeks between the first deadline and publication.
We’ll get into the details of each of these deadline stages in just a moment—but, in brief, they consist of:
- A research and outline stage, which writers use to conduct primary source research, determine the structure of their piece, and produce a detailed outline—which editors then evaluate to ensure the writers are heading in the right direction;
- A draft stage, during which time writers expand on their outlines, focusing on producing creative and engaging content;
- An editing stage, in which editors work with writers to identify opportunities to elevate drafts to a higher level and catch any hidden errors, and;
- A client review stage, during which time the client has an opportunity to weigh in from a content and compliance perspective prior to publication.
The bulk of the work takes place during the first stage—both because it makes the process much easier downstream, and because a heavy focus on research and meticulous outlining is a major differentiator in content marketing, particularly in the AI era.
Let’s take a look at that stage in a little more detail now, starting with research and structuring.
Step 3: Developing Your Concept Through Research and Structure
It’s worth noting that search engines like Google are agnostic about how a piece of content was created when determining how it ranks—instead, it’s all about how helpful and reliable the content is.
While (as of this writing) AI is incapable of producing such content on its own, as in other aspects of our work, AI tools can and do play a valuable role in creating the authoritative, trustworthy content that internet users want—and that search engines reward.
At TELL ME MORE®, this means writers take advantage of AI when conducting primary research and refining an article’s structure, in order to ultimately craft a robust outline—the starting point for a well-ranking piece of content.
Sure, AI makes some things easier. But when something is easy, the differentiator becomes, who is going to do the hard work that others aren’t doing? When it’s this easy to create content, there will inevitably be consequences; losing trust is a big one with AI. That’s what we have to work to build and maintain.
Conducting Primary Research
Have you ever encountered an intriguing statistic in an article, only to discover upon following the accompanying link that the cited source isn’t a primary one? Instead, you find yourself immersed in a loop of articles, each referencing and citing the same statistic without providing any actual, substantive source to substantiate the claim. It becomes a game of “telephone,” where the same information is passed along without any concrete foundation.
Producing trustworthy content takes more than pulling together a few external links; it involves significant research that tracks all citations back to primary sources—another area where our writers turn to AI to increase their efficiency.
Some of the AI research tools we use most frequently include Consensus and Elicit, which make the process of identifying, browsing, and organizing scholarly articles more efficient. Additionally, platforms such as Research Rabbit can allow you to easily create collections of research you consistently cite.
As always, a human touch is necessary to ensure any primary citations are relevant and high quality—something our writers have the latitude to do, since we have more time on our hands thanks to these tools making the research process smoother.
Confirmation bias is always a concern with [AI research tools]; ask for scholarly research indicating that cigarettes are good for you, and you’ll get just that. At this point it still takes a human to ask the right questions and then go deeper, doing things like investigating the authors of a paper—even one published in a scientific journal—to discover potential conflicts of interest that could introduce bias.
Incorporating Original Interviews and Images
In addition to research, primary source interviews are at the core of what makes TELL ME MORE® content unique. Whenever possible, our writers go to the source to get original input from domain authorities with direct, holistic knowledge of their subject matter.
Returning to our example of the article on ‘how is bacon made,’ those authorities might be farmers, producers, chefs, food scientists… anyone in a position to provide authoritative insight that can’t be replicated anywhere else.
Not only does this make content more trustworthy for readers, it also makes it more authoritative, which can lead to other websites linking back to our content—boosting our expertise, along with our rankings.
Regardless of whether it’s possible to secure an interview for a given piece, writers also pull insights from a constantly expanding database of past interviews and written testimonials, ensuring that, no matter what, each and every article includes genuine insights from real authorities.
Aside from conducting interviews and gathering testimonials, we also make an effort to source high quality original images whenever possible—for example, from interview subjects, client customers, or other relevant sources. Real images can be more effective than stock or AI-generated ones in certain instances, lending additional credence to our content—which is why we make this a priority.
As it exists right now, pure AI content has sharp limitations when it comes to factual accuracy, both in terms of interpreting information and authenticating sources. This is why it’s still so important for humans to do research and validate sources. AI-generated pieces are perfectly content citing other unverified AI-generated pieces; we aren’t.
Honing Article Structure
The structure of an article typically emerges during the research process. However, even after a writer has pulled together some research, we’ve found there’s value in using ChatGPT to suggest a structure based on an article’s basic premise.
This is illustrated in the image below, which aims to solicit a response based on information on the brand, topic, and target audiences.
Why do this? Because it quickly provides a sort of external check on any assumptions and conclusions the writer may have made up to that point. Are there any gaps in their understanding of the topic? Does the response provide any inspiration for how to best present their findings?
Of course, writers can be much more specific with their prompts as well, seeking input on specific topics to include, title and heading ideas, questions a reader may have, and so on. The more texture you feed into the model, the better the results will be—meaning they’ll be more helpful as the writer makes their ultimate decisions about how to structure their piece.
Step 4: Preparing Your Outline
With a body of research and a solid structure in hand, our writers will next move on to composition, which begins with writing a detailed outline. We have a standard template for this, which we work on in Google Docs—simply the best option for collaborative writing and editing.
This template includes several reference fields that serve as an aid during the writing, helping writers stay focused on an article’s fundamentals. These include:
- The primary target keyword, which helps the writer stay oriented around the core goal of the article.
- The meta description, generated by the writer, which not only serves an SEO purpose but also helps the writer hone their message by thinking about the piece from the perspective of a reader deciding whether or not to click through.
- The article objectives, which elaborate on the strategic purpose and context of the article, and include secondary keywords to target.
- The target personas, which refer to internally defined audiences with specific attributes that writers attempt to appeal to.
Why We Outline
The goal of the outline stage is to flesh out the structure into a highly detailed article roadmap. Specifically, it spells out (in detailed bullet points) the entire narrative arc of the piece. The language isn’t polished and the details may not all be present yet—but the outline nonetheless tells the entire story, broken out into headings and subheadings to make it easy for the editor to quickly recognize and assess where the piece is headed.
The outline also includes quotes from interview subjects (or from our database) and placeholders where images may go—though actual image selection doesn’t happen until the next stage of the process.
Upon completion, the writer passes their outline off to the editor, who makes an initial pass with the goal of assessing the big picture. Is the article’s structure coherent? Does anything seem to be missing, or is anything extraneous present? Do the transitions from one section to the next make sense? These are the type of questions they’re asking at this point, in addition to making suggestions with regard to tone, style, and conceptual approach.
Some writers are a bit thrown at first by the practice of writing such comprehensive outlines prior to the draft, but it really does save time—on both the writer’s and the editor’s sides. It helps the writer focus their later efforts on making the copy more engaging to read while conveying authority and trustworthiness, while allowing the editor to intervene early to make sure the article stays on track.
Step 5: Producing Your Draft
After this first editorial round, the outline heads back to the writer so they can spin it out into a full draft. This stage tends to go more quickly, since all the strategic planning has already been done, and the content and structure have been signed off by the editorial team.
AI’s Role in the Composition Process
It feels like it’s been a while since we talked about the role of AI, and it wouldn’t be surprising if the reader was wondering whether or not it plays a role in this stage. After all, generating coherent prose quickly is ChatGPT’s strongest suit; if saving time is the name of the game, wouldn’t its input be welcome here?
The answer is, yes and no.
When it comes to actually generating content, the short answer is, no. Generally speaking, the longer the content in question, the more cumbersome it can beto use AI for composition—because the copy will inevitably need cleaning up, and with lengthier pieces it will take as much (or more) work to polish it than it would to just write it in the first place.
However, it can be useful in vetting your own work after you’ve gone ahead and written it. Writers can pull individual sections—or even the piece in its entirety—and ask ChatGPT to analyze it for opportunities to make improvements. For example, you could ask, How could I reword this introduction to make it more engaging? What are some ideas for examples I could include to illustrate this point better?
Using AI like this is an effective way to bolster the quality of your writing without getting sidetracked trying to rein in AI’s often off-base attempts at long-form writing.
In addition to hammering out the text, writers are also responsible for putting the finishing touches on all other aspects of the piece—like selecting all photos, writing image alt text, and ensuring all H2s and H3s are formatted correctly.
This takes a little extra work on the writers’ part, but we’ve found it saves time downstream and is ultimately the most effective way to work. It takes much less time for the writer, who’s already embedded in the piece, to knock out these final details than it does for other team members to attempt to resolve those points themselves ahead of publication.
Step 6: Editing Your Piece
The draft stage culminates in another round of editorial review—which, this time, is aimed at refining the piece to the best version of itself prior to sharing it with the client.
Scoring the Content
This editorial pass includes the use of an article scorecard we’ve developed, which is based on Google’s EEAT guidelines—content guidelines that essentially provide a roadmap for creating helpful, reliable, people-first content. That acronym stands for Experience-Expertise-Authoritativeness-Trustworthiness—four key aspects of high quality content.
The scorecard includes a number of questions adapted from those guidelines, such as:
- After reading your content, will someone leave feeling they’ve learned enough about a topic to help achieve their goal?
- Is the article structured so that it is reader friendly?
- Would you expect to see this content in or referenced as an authoritative source online or in print?
Our editors use this resource to evaluate each draft that comes across their desk to ensure every piece of content we produce meets our standard of quality—which, again, is not an abstract concept, but rather is based on criteria that actually matter to internet users.
Enhancing the Copy
Beyond this scorecard, editors seek out opportunities to make the content more engaging for a reader.
These opportunities may be structural, like creating stronger transitional language between sections, but they’re often more tactile. We encourage our writers to deeply engage their physical senses when drafting, and the editing process is often where we refine that balance between strong imagery a reader can feel and a consistent, helpful flow of information.
AI’s still pretty lousy when it tries to include genuinely engaging imagery in its prose. It’s one of the technology’s most alienating features to a reader, so it presents us an opportunity to connect with the human beings reading our content. It’s critical to care about those real people as much, if not more, than we care about bots crawling our pages for SEO.
Checking for Plagiarism
Finally, the last step the draft goes through before shipping off to the client is to pass it through a plagiarism checker. We use QueText, but there are plenty of options out there. Checking for plagiarism is simply a best practice that anyone publishing content online should observe.
Accidental plagiarism can happen, especially when working with a lot of primary sources; and now, we have to add a new AI wrinkle to that problem. Since everything AI knows has been learned from other pieces of text, it’s not outside the realm of possibility that reproducing any amount of AI-generated content could actually lead to inadvertent plagiarism. The fact is, we just don’t know because AI is pulling from many data sources. That’s why we check, and you should too.
Step 7: Bringing It Home: Client Revisions, Publication, and Monitoring
Once our editorial team has completed its second review, we send the piece to the client for another round of review and revision. Affording clients the opportunity to review the content before publication is important for maintaining alignment—and essential for ensuring compliance.
Once the client has completed their review, our editors fold in any outstanding comments and package the article for publication. After QAing the article to ensure it’s been published correctly, we shift to monitoring the article’s performance. For that, we turn back to Semrush.
Here, you can see that all our hard work paid off. Within a month of publication, our ‘how is bacon made’ article jumped to number one in the rankings for that keyword—in addition to ranking first for several other related keywords. That’s thousands of impressions a month, and a steady stream of visitors to our client’s website—all from a single blog article.
Putting the Client and Audience First, Using All Tools Available
If this process seemed lengthy and involved, it’s because all of these elements really are needed in order to consistently deliver high quality results for clients. And it’s not just the elements themselves, but what they represent as a whole—a holistic attempt at creating web content that human beings actually want to engage with.
You can get results by practicing any of these steps on their own—but you’ll really notice the shift when you embrace a mindset that puts the client and their audience first. By all means, use AI to try to do that. We’re going to keep trying, too. It’s a work in progress, and our approach is bound to continue to change as the available tools change as well.
Just keep in mind that AI is a tool, not an end in itself; the real goal is enabling your team to make the best decisions possible—decisions that help deliver on your client’s needs.
Ready to take your content marketing and SEO to the next level? Contact us today or schedule a meeting to explore how we can design a strategy tailored to your brand’s goals. Let’s create content that not only ranks but connects with the people who matter most.