AI marketing is a lot more than machine-generated content. Today’s artificial intelligence tools provide business intelligence: marketing strategy, advertising plans, conversion-boosting ideas, and more, all produced from data.
But garbage in = garbage out. So the way you use these tools is as important as adopting them. Let’s see what works.
- Adopting an AI tool is one thing – using it effectively is another
- There’s more than one type of AI marketing
- It’s not all content – there are AI tools for planning and strategy too
- Machine intelligence works better with human judgement
- There are far more AI tools in the market than ChatGPT
- Think of AI in marketing as a productive assistant, not a department head
Old-school advertising agencies – which up to the 2000s were the go-to for any company wanting to spread the story of its products – all shared the same basic way of working. Account executives would understand the client’s needs and write the brief to a creative team, usually a copywriter and art director. In tandem, a media planner would set out a strategy for where that creative content would appear: all the channels and publications of interest to the target audience. (Or at least the channels and publications that’d bribed the media guy with lunch the week before.)
While ad agencies have evolved – the biggest now run on data, with analysts and strategists outranking mere creatives – that basic structure now has a new life. Because it’s how the new generation of AI marketing tools works. A blend of strategic understanding, original content, and tactical decision-making that engages the target psychographic and follows their buying journey as it progresses down the sales funnel.
Businesses and technologies change, but the basic way of getting things done hasn’t.
What we’ve found at Contentoo, though, is that getting the most out of AI marketing requires careful thought. In this blog, we’ll set out the different methods for AI assistance and offer some tips for putting AI marketing tools to work.
What is AI marketing?
AI marketing is any activity that contributes to your organisation’s brand image, customer engagement, and sales/profits by using AI software as a strategic tool.
It means chatbots trained to answer common customer queries without human involvement. It means software that watches clickthrough rates and tests one headline against another, or one offer, or one CTA. It means machine learning that sorts and sifts large volumes of data for subtle patterns that contain fresh opportunities.
In other words, the term covers a lot. To get a grip on this broad area, let’s look first at three typical applications for AI in digital marketing.
A few types of AI marketing solutions
The helicopter view first. Three approaches to artificial intelligence in marketing include machine learning, AI for Big Data/analytics, and AI as a marketing platform.
Machine learning is where AI comes closest to its “intelligent” label. Take a large dataset – even if it’s unstructured or non-textual, like a set of images – and create software that uses that data to “learn” relationships and patterns within the dataset.
It’s how “Captchas” work, helping machines recognise which images from a set contain a bicycle. (Yes, when you complete a captcha, you’re training an AI.) It’s how chatbots parse your questions and offer answers based on how similar your question is to one it’s heard before. And it’s how flavour–of-the-year ChatGPT works, applying statistical likelihood to a series of word-like “tokens” to generate content that sounds like natural language.
Machine learning uses a variety of methods known for years, including cellular automata, genetic algorithms, and neural networks. But it’s only in recent years that “compute” – the generic term for the power available from datacentres on demand – has been plentiful and cheap enough to make it viable for consumer applications.
Big Data and analytics
In 2020, the world created and consumed 64 zettabytes of data (Zb) (that’s 64 trillion gigabytes); five years earlier it was just 4Zb; by 2025 it’s forecast to hit 180. This is far more than any human could make sense of. Fortunately, if there’s one thing machine learning loves more than data, it’s Big Data.
All analytics involves statistical prediction: if a customer of Type A had a good customer experience with their €400 purchase, they’ll spend €460 next time. With a few hundred examples, your marketing team can safely make this assumption. But if the dataset grows to 10 million customers, and you have thousands of data points about each, you can make far deeper and more confident predictions.
Customers of Type A add 20% to your profit margin for every 20 positive reviews on TrustPilot; a single day of delay in delivery leads to a 5% churn among customers in the Northeastern USA in blue-collar professions.
Analytics use of AI can find these patterns in your data, and come up with insights you can use to focus your marketing efforts on what’s profitable.
AI marketing platforms and tools
To build customer relationships, you need communication – whether that’s email newsletters, blogs on your website, or the tone of voice your chatbot uses. So, a third leg of AI is the platform you use to do it at scale.
No human can top-and-tail 1,000 emails with a slightly different paragraph, or mix a set of offers to give every recipient the 3 they’re most likely to buy. But an AI assisted marketing platform can. It can write using a template, discover best practices, customise and tailor every subject line and sentence according to rules it learns from what happened before.
That’s why the world’s top content marketing platforms – including Contentoo itself – are embedding AI deep in their business processes. Because it lets customers answer their customer demands, individually and at scale.
AI marketing use cases
The possibilities AI marketing tools open up make it hard to know where to start. So next, let’s look at some practical things you can do with these AI approaches, starting today.
While ChatGPT popularised content creation tools, there’s actually a wide range of similar, some of which have been around for years. This technology uses AI’s predictive text abilities (known as “Large Language Models”) to produce articles, blogs, and other content that sounds like natural language.
What matters here is curation: how you use tools like Jasper and ChatGPT. The low-end, low-cost content you previously went to Fiverr and Upwork for: product descriptions for a catalogue, generic how-tos to boost your site footprint, brochure blurbs that fills up space?
Yes, AI content generation is a great help – and can do a great deal of the work with minimal prompting. But ask them to produce a carefully-considered white paper, or a “thinkpiece” with a real creative concept, and they’ll fail. Because machine learning isn’t capable of that … yet.
Given this, Contentoo combines the power of creative talent and AI-powered, cutting-edge technology to ensure top-notch content creation.
AI marketing tools can shine when you need a content refresh: maybe your FAQs and product copy simply sound tired and out of date.
PropertyGuru, the leading property search platform in Southeast Asia, was facing a challenge. The company had an abundance of outdated legacy content. Given this, its SEO positions were eroding. If you are interested in how this problem was solved, check out our case study on how PropertyGuru refreshed content at scale using cutting-edge AI technology and creative human talent.
Asking a content generation AI to produce a different version of a piece of text, or to rewrite existing copy with specific guidelines for Tone of Voice, can lead to colourful, more engaging content that brings readers back. Of course, you can do this to your email marketing and online advertising too. Just make sure you A/B test your new variants to make sure they’re delivering.
Automated email campaigns
Another twist: instead of setting an outreach calendar for your email marketing campaigns and pressing “Send” at set times each month, why not automate them?
Many content marketing platforms are experimenting with this “on-demand marketing”, providing a customer with the right information at the most opportune moment rather than simply doing a monthly campaign. And many SaaS providers, from plug-in tools for websites to cloud-based AI services, are offering them.
Some people groan at having to type into a chat window, with bad memories of unresponsive agents and long wait times. AI marketing bots are changing that. Powered by machine learning content generation, they’re now able to answer fairly complex questions – and contribute to sales, too, by suggesting (and linking to) items the customer may find useful.
New chatbots integrating internal company datasets with external deep learning are finally fulfilling their promise, and customers are finding them genuinely useful.
Challenges for AI marketing
Of course, AI’s not a solution for everything – you’ll have recognised some problems. Above all, understand that artificial intelligence may be “intelligent”, but it’s not sentient. It’s simply responding to data; it has no inner life or capacity for abstract thinking, much as it sometimes seems to. Here are some things to watch out for.
Training time and data quality
Garbage in, garbage out. (GIGO has been in use as a hacker’s rueful complaint for decades.) Feed your AI bad training data, and its generated responses will be off. (Some call this the “Uncanny Valley”, where things look right to begin with, but a second glance shows there’s something … not quite right.)
Getting this right is a tough challenge, because it means anticipating situations you hadn’t considered. (You don’t know what you don’t know.)
So first, make sure your training dataset is matched to the output you want. If you’re using it to generate 10,000 catalogue blurbs, your previous 500 bits of content are a good start. But perhaps you want to apply a new style, or make the average sentence length shorter, or cut down all use of idiom or slang. AI is great for doing this – if you think critically about what you want.
Deployment of best practices
At Contentoo, we don’t recommend going all-in on Day One. Adopting AI marketing is more like a learning curve: plan, test, roll out, and assess what happened.
That’s why, when adopting AI, it’s best to make it a gradual process. Test with a single use case (perhaps an email campaign), check to make sure that the output matches your expectations, and note where it doesn’t. Then add more data, note the changes, rinse and repeat.
Most marketers realise this is why AI is there to assist them, not replace them. Content creation, and the curation of the results that AI produces, benefit from human insight and judgement.
Adapting to a changing marketing landscape
AI marketing provides an advantage – but not everyone’s taking advantage at the same time. This means that a key AI strategy is simply running fast enough to stay in the race.
But if everyone’s using ChatGPT for their content, won’t everyone’s content sound the same? Overwhelmingly, yes. Users of some content marketing platforms complain that a single brief gets answered with 200 nearly-identical proposals.
And as marketing departments become more risk-averse and generic in their ToV requirements, it’s already getting hard to tell brands apart.
It’s a problem. But it’s also a great opportunity. So let’s look next at how to use these tools in the most intelligent way.
How to use AI in marketing
Everything so far suggests that a little forward thinking goes a long way when you’re adopting AI marketing tools. So here’s a little checklist to get started:
- Work out what you want, with an integrated AI marketing strategy.
- Do your homework, by researching the various AI marketing tools out there.
- Select the one (or few) that best answer your needs.
- And start small with a test or two before rolling out enterprise-wide!
To finish up, let’s look at each stage.
Create an integrated AI marketing strategy
A good start: look at your current marketing strategy, and see where AI could help. Whatever you do, don’t have an “AI marketing strategy” separate to your main one.
Is your product mix full of high-value, big-ticket items, like luxury houses or private jets? Each sale needs far more human contact and personal service than, say, buying plastic tableware. So AI isn’t the right choice for customer service here. But for ongoing customer contact, it may be a great option, able to select and customise up-sell and cross-sell offers tailored to each individual customer. (At scale.)
Same for lower-ticket items. Even if you sell a single range of kitchen implements, customers may respond differently to settings, colourways, even the race and gender of people in promotional photos – and AI can mix and match these elements in ways that maximise the selling opportunity for every fork and spoon.
For an AI marketing strategy that works, make it part of your existing one.
Research examples of AI in marketing
There’s no substitute for good case studies. So next, look for real-life examples of what AI’s doing for customers in your sector. Is a picture emerging of what AI does well … and what it doesn’t?
Some brands are finding “less is more”: fewer communications on fewer channels create an impression of exclusivity. Others are successfully rolling out small-scale customer strategies to larger audiences with the help of AI, increasing their value for no increase in cost. Conversely, some are discovering where not to use AI: perhaps top-value customers who want a human face from their vendor.
Whatever business you’re in, try to get an idea of what your competitors are actually doing with AI marketing tools.
Survey the different AI platforms
Last, with so many AI assistants out there, it’s hard to imagine which one to choose. The basic strategy here: choose a platform that lets you integrate different apps and tools, to help future-proof your AI marketing strategy.
A platform provides a base for sharing data between applications (for example, a customer database and your email marketing hub); for creating content and publishing it to the web; even collaboration between translators and project teams where tasks can be assigned and approvals gathered before your AI assistants go into action.
Obviously, we have some ideas about the best one. But the decision is yours.
Benefits of leveraging AI in marketing
Let’s end with some benefits. Here’s a blow-by-blow of what you can expect with smart use of these smart tools.
Increased campaign ROI
As every marketer knows, improving campaign performance isn’t a sudden step-change, but a gradual process of listening to data, making small adjustments, and repeating the positives over time. AI analytics is very good at measuring these metrics.
So don’t ignore that tiny uplift of 0.012% on your last campaign, or those two customers among 1,000 who bought an extra SKU on their last site visit. Over time, all those little wins add up to big results.
Better customer relationships, real-time personalisation
If you imagine the customer relationship as a managed conversation over time, the frequency and content of those communications – including the channels they happen on – are important. AI marketing tools can help you reach out to customers with the right choice of words, on the right channels, at the right time. This all happens with the goal of increasing P2B (Propensity to Buy) while increasing customer satisfaction and engagement.
There’s another benefit. In the early days of the web, the low cost of email marketing meant the customers you knew most about were often deluged with too many offers and messages. (Why tailor your offers, if it costs nothing to send one more email?)
This is now recognised as a bad strategy. AI marketing tools can plan the correct number of communications to maximise the relationship, without annoying the customer with too many.
Enhanced marketing metrics and measures
All data is good data, but what’s the best data for humans to look at when there are only 24 hours in a day? It’s the right set of metrics and measures.
AI can mine data for interesting patterns and insights, and summarise them on your marketing dashboards. But the more innovative tools are also coming up with their own improved metrics – findings with particular relevance to your business, making use of broader datasets.
For example: you already knew your winter coats sell well in autumn … but market data suggests you’d make more money if you sold them in August, when customers are booking their next ski-ing holiday. New metrics, new opportunities.
Faster, smarter, better decision-making
Last up comes the core value proposition of artificial intelligence in marketing: decision support for management. It’s sometimes said the main challenge of management is to make decisions based on incomplete information; with AI, that incompleteness becomes a slightly smaller problem.
You won’t lose your job to an AI. But you may lose it to someone who’s using AI better than you. Don’t be that version of you.
Your C-Suite will be onboard your latest strategy faster if you show them how a change will increase sales by 5.5% next year. Or how a rejigged marketing mix will double your margins at the same budgeted cost. Or that a campaign to a new psychographic segment has an 80% chance of making an extra million in sales before Q4.
AI marketing tools give you this backup. And ultimately, step by step, they’ll help you become more and more successful in your own job.
Conclusion: Start small and scale fast with Contentoo AI
So there you have it: the upsides (and downsides) of AI for marketing. Make no mistake, it’s not easy. But it is exciting. And if you’ve got big ideas for your business, there’s a long list of AI marketing tools out there ready to help you.
One of them, of course, is Contentoo’s integrated AI content marketing platform, blending the best of human freelance talent with the most carefully thought-out AI tools. There are still a few slots on our Beta program – why not come and test us out?