It’s a common scenario in marketing: you’ve created a complex new asset that has taken weeks, if not months, to nurture to the point of publication. Now it’s hot off the press and on your website, and you wonder what else you can do to raise its visibility.
One of our clients recently put this challenge to the combined forces of our Editorial Strategy and Visuals teams for a newly launched market research report.
Of course, we’re seasoned brainstormers.
But this time – despite our creatives chomping at the bit – we didn’t want to rely on our chops alone.
To ensure no stone was left unturned, we called on some of the generative artificial intelligence (GenAI) tools that have become part and parcel of how we work at Formative.
Could they “supercharge” our typical marketing ideation process?
Laying the foundations for content ideation with AI transcriptions
In our industry, like many others, generative AI is eliminating repeatable, rote tasks freeing us up to focus our efforts on strategic and creative thinking or more in-depth research, for example.
Note-taking is one of those repetitive, time-consuming tasks. That’s why transcription tools were some of the earliest forms of AI we used. Only a few years ago, transcription still relied on humans and came at a high cost. Since then, AI has democratised access, and today we use live transcripts from the likes of Otter, Trint and Google Meet almost every day.
For our client brainstorming and many other meetings, rather than waiting hours for someone to write up their notes, live transcripts provide detailed written records within seconds, as well as summaries of the key points. This means we can focus more of our time on developing compelling ideas.
Speeding up marketing ideation and management processes
While transcription tools offer summaries and key takeaways, we also harness large language models (LLMs) like Claude, Perplexity and ChatGPT to help refine and generate comprehensive meeting summaries and action notes.
With the right prompting, these advanced AI models can be very reliable, reducing the need for manual tweaking. Even so, it's crucial to exercise caution when using open-source LLMs, as they may pose risks to confidential information.
Paid tools, including Frase, ringfence and protect confidential content, ensuring data privacy and security.
Collaborating with AI for human-machine content ideation
During the client brainstorming session, we presented an initial list of ideas to promote a new market research report.
But we didn’t leave it at that.
We prompted Claude AI to come up with alternative angles to add to the mix, which ensured we covered all possible bases.
Incorporating Claude, ChatGPT, Perplexity and other LLMs into our editorial processes is becoming second nature. They support us with everything from accelerating and broadening our desk research to identifying content gaps and supporting SEO keyword research.
Naturally, AI has its limitations. Defined by its training, it can only extrapolate from given parameters and sources, and all outputs must be thoroughly fact-checked. But taking this “co-piloting” approach has shown us that AI can be helpful at all stages of ideation and, as in this case, turn ideas into content campaigns more quickly.
A human future that incorporates AI
Combining the power of GenAI tools with our decades of experience as writers, editors, designers, animators and strategists delivers the best results.
While we follow strict guidelines to avoid the pitfalls of the tech and to be transparent about how and where we use it in our work, we’re also keen to experiment. We relish exploring the benefits of GenAI to enhance our creative processes and keep us ahead of the curve as its use and capabilities continue to grow.
No single tool can do everything, so experts must pick the best combination for a particular task. Success or failure is largely determined by the sophistication of the prompting – essentially human expertise.
As many marketing organisations explore how to streamline their operations with the help of GenAI, it’s this collaborative human-machine approach that is likely to win out.
Author:
Andrea Willige
