How We Use AI at 1827 Marketing (And Why It Matters)

AI for marketing is the next great frontier for brands and advertisers. However, many in the industry are apprehensive or sceptical. Some can see its potential. Others see it as a threat to job security.

No matter your perspective, one thing is undeniable: AI is here to stay, so you need to equip yourself with the necessary knowledge to use it to your advantage.

We wanted to share how we’re thinking about and using AI at 1827 Marketing. We hope that our perspective and practices can help other people navigate their own decision-making on AI.

How 1827 Marketing Approaches AI for Marketing

First, we believe anyone leveraging AI should be transparent and open about how they use it. While we have always been enthusiastic early adopters of marketing technology, we've never been recklessly so. It is right to be cautious and considered in this rapidly shifting landscape.

In some ways, the recent shifts in AI reminds us of the early days of social media marketing. That too was a major change in the industry, requiring marketers to adapt their skills rapidly to stay relevant. This recent change might feel more intense due to its speed, but we're cautiously optimistic.

We're happy to use established and proven AI and machine learning in production for our customers and in our own work.

When it comes to more experimental AI, we don't use untested tech in production. Behind the scenes, however, we're constantly experimenting to stay on top of changes and to work out what we can confidently recommend.

Regardless, the fact remains that creating high-quality content that connects with audiences requires creativity, insight, solid research, and fact-checking. That means human writers, editors, and strategists will always be the beating heart of our operation. The content that we produce for our clients and ourselves is produced by a team of carefully chosen, experienced and talented human writers and designers.

For us, AI is like a diligent and extremely well-read intern. It helps us to optimise our workflow and allows our experts to focus on the tasks that only they can handle.

Does our new AI team member have limitations? Absolutely. But we're finding that with some tight supervision and direction in place, they're becoming more indispensable by the day. We’re excited about their future potential.

Accelerating Changes in AI

If you find the rapidly changing landscape challenging, you’re not alone. Everything is moving fast and no one knows how these developments will shake out. 

The truth is AI is already embedded in your processes. No matter your perspective, there’s no need to be intimidated by the accelerated changes — but you must stay educated! 

New AI software hits the marketplace constantly. ChatGPT might have been first out of the gate, but AI-assistants like Google Bard, Claude, and Poe aren't far behind.

Several capabilities people are just now getting their heads around are already old news. ChatGPT is already being replaced by autonomous agents like AutoGPT. Image generation is making leaps and bounds towards becoming more conversational. 

It's also clear that AI will become a ubiquitous integration to all software. Brands like Notion, ClickUp, Canva, Adobe Creative Cloud, and IFTTT have already launched their AI offers. Microsoft is incorporating Copilot across Office 365. 

In the time it took to write and review this article, announcements came thick and fast. Google announced it was bringing generative AI to search. And Squarespace and Elementor for Wordpress both announced new AI-integrated offerings. 

How AI Fits in Our Current Workflows 

AI is in the news more than ever, but that doesn’t mean it’s new. Machine-assisted marketing existed when we founded 1827 Marketing and it has been part of our toolkit since day one.

Ideation and Optimising the Brief

Sometimes the most intimidating part of content creation is the blank page. So, that’s where we start putting AI to work. It can be a great partner for brainstorming, and it can help us to crunch through large amounts of search data to identify important themes and trends.

The best way to come up with new content ideas is often with an audit of existing successful content. AI can examine current titles and offer suggestions for other content that your readers will find helpful or engaging.

We also use AI to generate ideas through keyword research and clustering. It helps us discern what people are searching for, organise topics, and group them by meaning. This builds SEO in up front in our content planning and briefing process. AI support, through tools like SEMRush, also enhances our ability to create clear and effective SEO guidelines for our team of writers.

A further use for AI is in cutting the time required to stay up to date with trends. Software like Notion, Feedly, and Readwise's Reader RSS reader and can be used to filter and summarise articles for both briefing and content curation.

Good content has to be comprehensive, expert, authoritative and trustworthy. It also has to be competitive. The systems and tools we have assembled to process and sift large volumes of search and market data help our writers to address topics more thoroughly .

We don’t just use off-the-shelf tools either. We’ve even built and trained our own AI models to generate and evaluate ideas for a client's brand naming project, allowing us to create and consider many more options than is humanly possible.

Review and Approval

Once we have draft content from our human writers. We use AI to analyse text for SEO, readability, clarity, style, and grammar.

Our writers and editors use AI enabled services like Grammarly and Pro-Writing Aid to analyse their copy before submission.

Before publishing, we use SEMRush's SEO Writing Assistant to get immediate suggestions on optimising a piece. This includes a very handy AI rephrasing tool.  

We also have software to review content to help guard against plagiarism.  

In the same way, we now use AI-checking software to ensure written content isn’t authored by machines. This helps us ensure integrity in our work and great content for both readers and search engines. 

Content Distribution and Promotion 

Once new content is on our site, AI-powered amplification tools automatically create social media posts based on the copy. It suggests captions and images that we then review, edit, and optimise. This, for us, is a great use of AI to augment and accelerate our processes but not to replace them. The machines make a first pass and make suggestions that we can consider, but in the end an experienced human editor and social media manager considers and recrafts the AI’s first draft of social content.

We take full advantage of programmatic advertising options on major PPC platforms like Google, Meta, and LinkedIn. This improves ad efficiency and reach but requires human supervision to approve ad placements and ensure the desired ROI. Predictive analytics help optimise the ad and its copy for the target audience before it ever even goes live. 

Data Cleaning and Web Development 

Agencies like ours deal with a lot of messily structured data and CSV files that we might generate or which our clients might share with us. We need quick and repeatable methods to take raw data and extract what we need for it. Here, we use AI to assist us with developing formulas and Regular Expressions (RegEx) to parse the info on our behalf. For example, when auditing a brand’s content, we may start by extracting data from their sitemap. An AI like ChatGPT can be an invaluable assistant as part of the process to process the map and pull out just the URLs for the articles we need to review. 

When we’re developing custom automations or website features, AI can also assist with coding languages like Python, CSS, HTML, and Javascript. It provides a more accurate shortcut to Googling the answer for tasks where a developer would be overkill. Likewise, in programs like Excel or Notion, it can help you to create formulas to interogate and display your data.  

Where We’re Experimenting With AI 

As new AI tools emerge, we continue experimenting with the technology to find out what is possible. 

Large Language Models (LLMs)  

We do not use, or recommend using, AI to generate long-form content from scratch. We have a hard time imagining a world where that will change.

AI's limitations are exposed when humans aren’t part of the equation. Still, it provides a foundation that, at the bare minimum, makes for a great brainstorming partner and can lead to creative breakthroughs.

AI can be quite adept at assisting with short-form content. As a content creator, it's great for getting you off the starting blocks for things like product descriptions, social captions, ad copy, and email marketing subject lines.

Quite often, we use an AI like ChatGPT to experiment with different tones of voice and personas. If we’re in the process of creating 300 alternative headlines and descriptions for a Google Ads campaign, getting an AI to be more playful by asking it to try out different perspectives and styles - even asking it to generate verse or experiment with literary genres - can help our team to unlock their own creativity and find a broader and fresher set of approaches to producing the final copy. Similarly, we might ask an AI to critically evaluate some advertising copy from the perspective of our target persona, and this can help to sharpen our thinking.

Image and Video Creation 

Another area of experimentation is with image-generation tools like Midjourney. AI image generation is improving rapidly, but it still has some very obvious limitations.

Have you ever noticed how hard is it to draw a human hand? AI bots have the same issue. At least the camera never lies when it comes to making sure the models in your brand imagery have the correct number of limbs or fingers. 

Nonetheless we continue to experiment with creating possible brand imagery to replace stock imagery for our content. Midjourney has a feature (‘describe’) that can take an image and reverse engineer prompts that might create something like it. Editing these descriptions and adding brand colours and brand styles can yield interesting results, but the limitations in reliably creating anatomy stop us from trying these techniques in production.

For once, all the images we have used in this article were created using AI, with prompts that encouraged use of our own brand palette. Some of them aren’t bad, but none of them quite bear close inspection. This approach isn’t client-ready yet, but we continue to experiment.

We’ve explored AI video capabilities to provide a “talking head” to narrate scripts in auto-generated social media videos. While impressive, as with most other complex creative tasks, there can be something a bit “uncanny valley” about the final AI product.

Our writers produce scripts optimised for human speech, but current AI avatars often don't reflect the way a person would actually talk. The emphasis placed on words or facial expressions can be off. But, like all other AI technologies, we're sure major improvements aren't far away.

Chatbots and AI Assistants 

Most recently, we’ve looked at AI-powered chatbots using ChatGPT and LangChain software. We found it came with unexpected brand safety concerns.

For example, we tested some commercially available products that create chatbots to let people talk with the content of a website. The expectation is, we think, that people would ask simple and short questions. But we found we could ask those same tools to produce sample blog posts and even landing page copy for anyone wishing to create a rival product with similar functionality. It used the data on the test site to provide marketing insights, content and even criticisms that would be incredibly useful for competitive analysis and rival start-ups.

Although these chatbots aren’t a good fit for public facing applications until those concerns are addressed, we see the potential of these resources in a closed environment such as a client portal or staff intranet. AIs that can interrogate PDFs and explain complex and lengthy documents, internal policies and previous proposals in a conversational way could be a valuable resource. 

Looking Ahead 

The future is AI. It’s naive to pretend otherwise. We have already written at length on whether ChatGPT Support, Transform, or Undermine Content Marketing.

Will there be major marketing job losses as a result of AI? Quite possibly. But there are also major opportunities. The question is whether you're able to see the potential and update your skills. Some people will be empowered to focus on the strategic and creative aspects of the job.

Most of the careers in digital marketing existing today were impossible to imagine 20 years ago. The same is true for AI. Roles will evolve, responsibilities will shift, but there will always be a need for human content marketers. 

AI helps our writers, editors, and strategists to focus on the parts of the job only they can do. It can do a lot, but even with the rapid developments in the field, we believe we will always need human experience, insights, and creativity. 

And the future of AI in marketing reaches beyond just content development and marketing automation. We see the potential for using agents to manage marketing campaigns, for real-time suggestions based on predictive analysis.

So, in summary, what’s our advice? Remain enthusiastic about understanding AI technology. Push yourself and your staff to learn and find new ways to leverage the technology for efficiency and optimisation.

But, remain ethical. Be transparent with customers and clients. Be on top of concerns around data privacy and IP, and act accordingly.

We are genuinely enthusiastic about the future of AI in marketing and content development. By using AI to enhance our workflows, we've amplified our effectiveness and productivity, all with the singular objective of propelling our clients' success.

Get in touch today and discover how we can expedite your journey to success.