A Practical Guide to Generative AI in Marketing
Our Recommendations for Tools and a Case Study With Sunday
The enthusiasm and speculation surrounding what generative AI could hold for the future remains steadfast. At the beginning of this year, reports showed ChatGPT surpassing [1 million users in 5 days](https://www.statista.com/chart/29174/time-to-one-million-users/#:~:text=Other popular online services have,via downloads of the app.) — 15x faster than Instagram, the next fastest online service to reach 1 million users. Momentum built throughout the first quarter with Microsoft reporting that they have more than 2,500 Azure OpenAI service customers, up 10x quarter-over-quarter and including Coursera, Grammarly, Mercedes-Benz, Shell, and Epic Systems.
While generative AI inspires any number of paths to business and product improvements, most practical applications are in exploratory stages. One of its most promising and least contentious use cases, and a category near to me and core to Forerunner.
For marketing teams, generative AI has already proved valuable in optimizing and accelerating creative work (from visuals to writing), offering routine workflow improvement with opportunities for output expansion. And this is all for an engaged group of professionals who are typically inclined to explore new technologies.
What We Learned
A recent Forerunner survey of over 100 marketers validated the activity we’ve been observing: marketers report clear awareness of AI and interest in using it to their advantage:
76% see AI as significantly changing their profession
63% believe adopting AI will make them look good at work
44% are “all in” on AI applications in marketing
25% are actively using AI applications today
Another 25% are taking meetings to start using AI applications
For context, compare the above to the HR industry: a recent Pew Research Center study found that over 70% of Americans oppose AI use in making final hiring decisions and over 50% oppose AI analysis in making firing decisions. Or last month’s open letter to “Pause Giant AI Experiments.”
The marketing-focused respondents to Forerunner’s survey, on the other hand, share a significant level of interest in AI for copy and images. When it comes to choosing vendors, they reported proven results and unique technology (such as, novel applications to be built on top of generative AI models, like ChatGPT and DALL-E) as the two most important factors.
However, cutting through the noise to choose a vendor presents challenges. The market has quickly become crowded with buzzy frontrunners, new entrants, and incumbents vying for distribution. And to add to that, key players appear to be converging toward offering all-in-one platforms, whether they started in copy, video, or images.
To offer perspective on how to assess marketing AI tools, we put five platforms to the test with creative assets from Sunday, a Forerunner portfolio company using generative AI for copywriting while spending significant marketing dollars during the spring selling season. The five platforms selected include a mix of companies with high awareness among survey respondents and a variety of product starting points (copy, images, ads, etc.) all with self-service functionality; they are meant to be illustrative of the platforms in a market that will no doubt continue to move fast. We spent a similar amount of time in each platform and uploaded the same creative assets where applicable with an eye toward price, creative output, and product offering and usability.
What We Found
The headline takeaway from our efforts: brands keen to use AI should prioritize starting in copy writing across paid search, SEO, social content, and paid social. The copy writing offerings are the most advanced and provide a compelling AI entry point for organizations. Once you’ve mastered copy, creating digital ads with visuals would be a natural next step. Those platforms, however, are less sophisticated today.
The table below outlines our findings and includes an example of a creative asset each platform generated where applicable. Across copy-first and ad-first platforms, cost doesn’t seem to be a meaningful factor — for a fast growing brand with a sizable marketing budget, monthly price points are relatively similar and the more expensive options are affordable relative to the value they create for teams.
Our TL;DR Recommendation for Tools:
Start trialing with OpenAI’s tools to see where AI can fit into your team’s workflow. Looking for caption or paid search solution after that? Go for Copy.ai. Looking for longer form content for a blog or SEO? Jasper and Writesonic are for you. Of the platforms we tested, Pencil will be best for brands with in-house creative teams (though don’t sleep on Treat and Minta if you’re able to get access). If you have a smaller team with limited creative resources, check out Simplified. If you’d like to generate creative visuals from prompts (rather than use your existing images), Runway, Midjourney or DALL-E are good options.
And now for our more detailed rundown…
Copy.ai, Jasper, and Writesonic started in copy. According to survey respondents in Forerunner research, Copy.ai and Jasper have the highest awareness among platforms. While each offers a range of copy features, Copy.ai excels at short form for ads, product pages and social captions, while Jasper stands out for human-like long form copy and Writesonic shines for templates and long form.
Copy.ai and Writesonic take less time to learn, while Jasper has a steeper learning curve. Even though all three offer the ability to generate images, the feature is early — there’s opportunity for a more remarkable experience that builds on the table stakes image generation of DALL-E. If they want to become all-in-one platforms, these companies will need to evolve in order to meet the needs of sophisticated marketers who want to leverage brand assets and create visual ads.
Writer, Peppercontent, and Copysmith are three other copy-first platforms we found in what might be the most competitive marketing AI category. Writer has a compelling brand style guide feature that ensures teams across a company are aligned on brand voice. Peppercontent and Copysmith are both focused on site content for now.
And then there’s Pencil, which started and specializes in digital ad creation with a robust ability to make static and video creative using brand assets. The platform offers a performance prediction software based on the KPI and audience a marketer targets. The prediction offers a gauge for whether or not Pencil believes the creative will be successful if launched — while this is compelling, most of the AI-generated ads Pencil created for Sunday were predicted as “losers” by the platform.
Treat and Minta are two other ad-first solutions we considered. Treat is not self-service at this time and Minta requires an ecommerce store integration to log in. Both platforms could quickly catch up to Pencil.
Simplified, on the other hand, has prioritized creating a hub for brands to use templates across design, copy, ads, and video. The platform aims to be more of a workflow and collaboration tool, though the templates are less inspiring and design-forward. The offering seems better suited for smaller companies who do not have internal creative resources.
Typeface is an all-in-one platform that we are eager to try when launched. The site shows a design-oriented, user-friendly interface. The dropdown support for prompts shown on the homepage could be game-changing.
How To Get Started
Which platform you choose will depend on where you’re at in the journey and what you want to use the technology for.
If you’re early to feeling out how generative AI may work for your brand, give ChatGPT and DALL-E from OpenAI a go. They are the foundational technology behind many of the tools considered in the piece and offer easy, free, and high quality outputs.
The tension in adoption may come from who champions and owns the tools internally — will this be a creative-led initiative that augments routine work once foundational brand assets are approved or will this be a performance-led initiative that eliminates a creative bottleneck? At Sunday, the creative team owns the platforms and has found the most success with ChatGPT for where they are at. I suspect as the platforms get more sophisticated and easier to use, performance teams will value the unlock to do more on their own with creative’s guidance.