It’s official: precision marketing is the new performance marketing. And a fundamental tenet of precision marketing is using LLM generative AI like ChatGPT or Claude to work smarter, not harder. But I know you don’t need to be told to leverage AI in your daily workflows for the millionth time - you want to learn how to do it. This guide will provide tangible, actionable use cases for AI in your daily workflows - with examples and next steps. It’s time to grow and evolve your current workflows by working with AI, and it all starts with understanding how to properly communicate with these LLMs to improve their output and results.
Choosing the right chatbot
The heavy hitters are ChatGPT, Claude, and Perplexity. Here are the strengths of each:
ChatGPT: A high-level strategic partner that’s good at editing existing work like blog posts or slide decks, structured problem-solving, and testing theories and assumptions.
Claude: A thought partner that’s great for coming up with new ideas, providing feedback, and pointing out potential risks and flaws in your strategy.
Perplexity: Can be treated like a research assistant. Great at gathering and analyzing data and fact-checking with access to real-time info, whereas other chatbots only have access to data up to their last updated date.
I have ChatGPT Plus, and I love its library of specialized GPTs. You can find a GPT trained to do almost anything here:
The value of using one of these specialized GPTs is that it’s already trained to do what you need it to. You only need to feed it contextual info about your business; then you’re off to the races.
Since ChatGPT is my AI of choice, I’ll use it to provide examples throughout this article. But before we get into specific ideas for using generative AI in your daily workflows, let’s cover how to talk to your chatbot.
Training AI to be a strategic partner
Generative AI works great for two things: helping you make decisions and helping you deliver on those decisions. But to provide value, AI needs as much specific information as possible. For precision marketers, this information will usually include links to relevant pages on your website, past pieces of content you’ve written, past campaign metrics or analyses you’ve done, anything that will help the LLM narrow its focus and define the playing field it should operate within.
If you’re starting from scratch, you’ll want to give the chatbot some framing of who you are and what you do. Here are some good bits of context to give the chatbot before you start sharing your prompts:
Who are you, and what do you do?
- What’s your role/mandate within your organization?
- What’s your company industry, size, and target customer?
- What types of challenges are you usually faced with?
- A link to your website
An example of a contextual note to share with your chatbot before prompting is:
I’m a social media manager for a shoe brand. Here is our website: [link]. Our primary customer is women ages 18-40. We’re a smaller, independent brand and often struggle to compete against bigger brands, but our unique designs have helped us develop a loyal following. Our pricing is mid-tier, with the average pair of shoes costing around $200.
The next step is to frame your goal with a simple, effective prompt formula: decision + outcome + constraints = happy marketer.
Start by telling the AI about the decision you need help making or the challenge you’re trying to solve. Next, share your desired outcome - what goal or measurable output are you hoping to achieve? And finally, share any constraints - what rules, limitations, or external factors does the AI need to consider before making a decision? Here are some examples of how to build a decision, outcome, & constraint (DOC) prompt:
Bringing it all together, your prompt might look something like this:
I want you to help me decide my next whitepaper topic. I’m trying to drive C-suite executives in the automotive industry to my website and entice them to download the whitepaper. Here is a list of topics we’ve covered in the past and a list of topics I’m considering: [insert lists]
This simple formula helps AI understand not just the basics but also your motivations and goals. This way, you can have a strategic conversation with AI, asking it questions and providing feedback on its responses (vs. asking it to output generic SEO-optimized title ideas for your blog posts). Using AI as a strategic partner means asking it to build an entire strategy for you, not just to complete tasks.
Many social media marketers use AI to draft copy for their posts. They’ll make one-off requests, like “I want to post about Earth Day. Can you write the copy?” Generic requests beget generic responses. And can you blame the poor chatbot? It doesn’t have any context!
The precision marketer uses AI differently. Rather than ask for one post, the precision marketer tells ChatGPT:
I’m a social media manager for [brand], and I want to increase leads driven from organic social media posts by 20% this quarter. Some of the topics I typically post about are fashion, lifestyle, and music. I like to post 3-5 times per week, make sure my posts are engaging, and encourage readers to share and leave comments. Can you please (the bots like it when you say please!) build a social calendar for me for the next month and draft the posts?
It takes some time, but sharing those details? That’s how you get AI to really revolutionize your marketing program.
Now that we’ve covered strategic prompting let's get into some specific ways marketers have seen success using Chatbots to streamline their workflows:
1. Draft thought leadership content
This one may seem obvious, and there’s a reason for that. AI does a great job at creating a first draft for new content! But to get the chatbot to write in your voice, you need to feed it some of the content you’ve written in the past. If you want the chatbot to generate LinkedIn posts for your brand, share a link to your profile so it can scan past posts and copy your tone. Or, if you want it to draft a blog post for you, share links to your 20 best posts so the chatbot can copy your tone of voice and gain background knowledge about your industry. Here’s an example of how to prompt your chatbot to draft a blog post using the decision, outcome, and constraints model we covered earlier:
I want you to help me draft a post for Marin Software’s blog about the future of digital marketing. I’m trying to spread helpful information and get the word out about Marin. I’d like this blog post to be SEO optimized so it will get at least 500 clicks. Some topics I’m considering include AI for digital marketers, breaking down ad-tech walled gardens, and a guide to LinkedIn advertising. I’m open to other ideas if you have any suggestions. Please copy the tone and structure of these past blog posts: [links here]
Here’s the result this prompt gets me on ChatGPT:
AI for digital marketers certainly sounds like a good topic for me to write about! 😉
As you can see, ChatGPT’s response is more like an outline than an actual blog post. Chatbots can get you started with a first draft, but you’ll probably need to build out the details and make the tone less robotic. Human intervention makes the content more readable and engaging and is also necessary for SEO.
Having generic AI-written content on your website can significantly negatively impact SEO since Google’s ranking system kills any content that it detects to have been written by AI. So, while ChatGPT gave us a great starting point, it’ll require a heavy editing hand before it’s ready to go live.
2. Get SEO-optimized product names and descriptions
SEO should be at the forefront of every naming decision you make. And thanks to generative AI, brainstorming has never been easier. Let’s say you sell women's clothes online. Rather than writing names and descriptions for each unique item, have chatGPT write them for you. Here’s an example prompt:
I want you to write a name and description for a new product on my website. It’s a floral women’s dress originally from Target. The title and description must be optimized for SEO and eye-catching to attract consumers. Feel free to give me a couple of options.
Here’s what ChatGPT gave me:
These options are too long and generic, so I asked chatGPT to shorten them and use trendier language. Here’s what I got:
Much better! This highlights a critical point—you have to work with the chatbot and give it feedback to get great results. Don’t give up if it doesn’t get it right on the first try. Keep telling it what to tweak, and eventually, you’ll get the content you’re looking for.
3. Social media strategy and post generation
I touched on this briefly in the intro, but let’s dive into the details. Many marketers use chatbots to draft social posts. ChatGPT Plus even has a handful of GPTs explicitly designed for this purpose:
There are also GPTs built to write copy or define your strategy for each unique social channel. If you have the paid version of ChatGPT, I recommend using one of these chatbots, as they are already trained on social media marketing best practices and channel-specific trends. Try out a few of the different bots by giving them all the same prompt and seeing which answer you like the best.
If you have the free version of your chatbot of choice, it’ll still work as a great social media assistant. Just make sure you give it tons of context; it may need a few more rounds of feedback before generating the perfect post.
As discussed earlier, we must use these chatbots as strategic partners rather than one-off content-generation machines to get the most out of them. Therefore, we should first ask the chatbot to generate a content calendar. Here’s the prompt I’d use:
I need you to generate a content calendar for LinkedIn for August. The goal is for our followers to engage with the posts. Our followers are mostly digital marketers. Each post should include a link to the most relevant landing page on marinsoftware.com.
The free version of chatGPT gave me a decent framework:
But I wanted something more customized and detailed, so I sent the same prompt to one of the social media post custom GPTs, and it replied with questions, wanting more details before creating the calendar and posts:
This is the sign of a great GPT! Asking for more context before providing output will lead to better post suggestions and less back and forth. This is just one of the many reasons the premium version of ChatGPT is worth paying for if you use it regularly.
4. Compare your website against competitors
If you want to understand your competition’s key value props, feed their website to a generative AI bot and ask it to summarize their strengths and compare them to yours. Here’s a prompt idea:
I need to understand my biggest competitor's key value propositions. Here is their website: [insert URL]. Please summarize their biggest selling points for me, then review my website: [insert URL] and let me know which of my competitor’s key value props are also represented on my website and which are missing.
I used a similar prompt to compare Google Ads to Bing Ads, and ChatGPT gave me a pretty good response:
5. Analyze competitors’ social media strategy
Did you know you can use Meta’s Ad Library to see all the ads your competitors have run on Facebook and Instagram? Simply select the ad type you’re interested in (I always do ‘all ads’), enter your competitor’s name, and hit search:
I’m using Target as an example. This search returns every ad Target has posted within a given date range that you can update:
You could then feed the resulting URL to your chatbot and ask it to summarize your competitor’s social strategy with a prompt like:
I need you to help me analyze my competitor Target’s social strategy. I’m trying to make sure my company’s social strategy covers the same value propositions that Target’s does. Here is a link to their Meta Ad Library results page: [link]. Can you analyze their posts and provide a summary of their strategy?
The result is a helpful overview of their paid social strategy:
The bot provided a detailed analysis and cited the sources it used to understand the most important elements of social media strategy. Nice!
6. Keyword research
Generative AI is great at helping you choose target keywords for SEO or paid search campaigns. Simply share your site with the chatbot and ask it what terms you should be bidding on with a prompt like:
I want you to generate a list of target keywords for my new Google Ads campaign. I’m targeting women ages 18-30 who make more than $50,000 a year and are interested in fashion. I sell shoes. Here is my website: https://www.shoedazzle.com/. What non-brand keywords should I be bidding on?
Here's the result:
And just like that, the chatbot has given us a great list of keywords to bid on. This list could be used for SEO, too – just incorporate the suggested keywords into your copy. You could even ask the chatbot to write product names and descriptions that include the target keywords or brainstorm topics for blog posts or other types of content that would feature the keywords.
7. Respond to comments and DMs
Social media community management can be tedious and time-consuming, so I recommend using a chatbot to auto-generate your responses to comments and direct messages. Provide the chatbot with context about your business and links to your social media profiles so it can analyze the tone your team typically uses when responding to comments. Then, input the comment you’re responding to and let the chatbot be your customer service agent! If the information needed to answer the question exists on your site, the bot may answer the entire question for you. If not, it’ll give you a template so you can fill in the blanks. Here’s an example prompt:
I need help responding to comments on my latest social media post. Here is a link to the post: [link]. I’m trying to respond to the comment, “How much does the software cost?” Can you provide a kind, professional, and concise response? All the information you need to answer the question should be on our website: [link]. Please leave placeholders for any details you cannot find, and I will fill them out.
Here’s what that prompt got me:
ChatGPT’s initial response was quite long for a comment, so I requested a shorter version and got this:
Short and to the point! Looks good to me.
8. Data analysis
AI chatbots are great with data. Simply feed the bot a dataset and let it work its magic! Let’s ask ChatGPT to analyze month-over-month trends in ROAS based on a dummy data set. The prompt here is quite simple since I’m just asking the bot to analyze numbers:
And here is the result:
It even gives me the option to download the table it created as a CSV! I could continue to ask further questions about the data, too, like what the average ROAS was for the year or the total cost.
This is just a basic example – chatbots can do much more complex math than this! For instance, you could ask the bot to analyze a much larger data set and estimate your ad budgets for this year. You just need to ensure the bot has all the necessary context to predict future trends. Make a mental list of all the data points you’d need for a predictive calculation. Then, ensure every bit of info on that list has been fed to your bot so it can do the analysis for you.
Other use cases
Those are just a few ways to start incorporating generative AI into your daily marketing workflows. Some other things you could use your AI assistant for are drafting emails, paid media campaign planning, brainstorming titles for blog posts, brainstorming names for new products or features, analyzing financial data, translating documents… the possibilities are endless! Just make sure you use a decision, outcome, and constraint (DOC) prompt, and a bot like ChatGPT should have the answers you need.
If you’re looking for a more customizable, AI-driven solution designed specifically for digital marketers, check out Marin. Our industry-leading AI takes paid media campaigns to the next level by fully automating the bidding and budget allocation process. Our platform also allows you to create objects and edit in bulk across all your different publisher accounts in one UI. To learn more about how Marin helps digital marketers harness the power of AI, click here.
