AI + Creativity news, resources and part 1 of a marketing toolkit
Leveraging AI to distill and unlock Business Issues
The future of creativity got a big, positive jolt this week. To frame it in a Jobs quote, how can we use spatial computing to “amplify our innate human abilities?”
Meanwhile, the world keeps shipping other opportunities, insights and possibilities.
It’s okay to feel overwhelmed.
You don’t need to keep up-to-speed with every single AI insight, hack or news. Honestly, I’m not either. Let’s all just agree there’s a lot to know, to comprehend, to try and we’ll just help each other along.
AI + Creativity News
NVIDIA, the dominant AI chip maker, is partnering with advertising holding company WPP to, “enable creative teams to produce high-quality commercial content faster, more efficiently and at scale while staying fully aligned with a client’s brand.” Is this yet another holding co gimmick? (Whatever happened to Publicis Group’s Marcel?) I watched NVIDIA founder and CEO Jensen Huang’s COMPUTEX keynote revealing the concept. In short…
1️⃣ The past was about the information constituting an advertisement being retrieved (and served) to you. In the future, your ad will be generated on the fly. The WPP demo starts here.
2️⃣ Digital twins (definition) are going to be much more commonplace. We already leverage this in CPG with packaging on Amazon’s “shelf,” et al. But tools like Nvidia’s Omniverse Cloud are accelerating the need and capability.
3️⃣ Then combine those twins with a tech stack including Adobe, Getty, Shutterstock—and WPP’s creative teams will generate the rules for live-rendered, licensable backgrounds, contexts, and scenarios in which products are pitched online. Think product landing pages, e-commerce workflows, digital ads. You won’t just concept and layout a singular ad, but choreograph a micro “world” in which each user’s experience could be unique (and branded).
In other news…
Bjørn Karmann has taken AI-generative techniques into the “real” world, with the Paragraphica, a device which uses GPS, direction, weather and an API to generate, “a scintigraphic representation of the description.” There’s a virtual version, of course. But you can see how ever-increasing pools of real time, and social imagery could fuel generative applications (i.e. Adobe Photoshop’s Generative Fill). And I love his star-nosed mole metaphor, “which lives and hunts underground, finds light useless.”
And finally…
As a thought experiment, ask yourself what tasks do you perform “where the final written output is important because it is a signal of the time spent on the task, and the thoughtfulness that went into it?” Professor Ethan Mollick is asking us to consider if that old fashioned signal disappears with The Button.
AI + Creativity Resources
A16Z’s AI Canon [May 2023] - The VC firm has done everyone a courtesy by evaluating and organizing a deep wealth of technical papers, and how-to resources. It leans towards coding and math, but overall is an excellent place to begin learning more.
To those links I’d add:
What every CEO should know about generative AI, via McKinsey [May 12, 2023]. Pro tip: This isn’t just for CEOs. The consultants provide a digestible summary, along with useful graphs and definitions. Maybe start here?
Google’s Generative AI learning path walks you, step by step, through the mechanics of LLMs, encoder-decoder architecture, and the “guts” of an elaborate process we’re all benefiting from.
LinkedIn Learning and Pinar Seyhan Demirdag have united to offer What Is Generative AI? a useful 42 minute course.
Or you could book one hour with me for Generatively Better.
11 Books That Will Help You Navigate the New Age Of AI, by Greg Satell [May 29, 2023]. You’re not going to read any of them. But you will appreciate Satell’s cogent summaries, which taken collectively, helps explain how we got to where we are.
What is Code? by Paul Ford, via Bloomberg [June 11, 2015]. Written eight years ago and still very relevant. If you’ve resisted understanding the terminology of computation, the world of coders, the logic of computer engineering—please block a few hours for Paul Ford’s story. This one article took over an entire issue of Bloomberg magazine. And for good reason.
Oh, and can we copyright the content we create using generative AI? The short answer is “no,” but there’s so much worth evaluating here and Descript offers a useful primer.
And it remains crucial to continue to ask “What CAN’T AI do?” as Dr. Brandeis Marshall puts it. Resolving conflicts, thinking critically and emotionally are tops on my list.
Part 1: Leveraging generative AI to help distill business issue(s) into actionable marketing
Today’s post inaugurates a multi-part unpacking. I want to use this “How Marketing Functions” framework 👆🏽 to dissect how you might leverage AI within each of the major steps in Marketing to improve strategies and creativity.
BUSINESS ISSUE(S)
Why does anyone care? i.e. Define the tactical problem, long term challenge, or opportunity which we believe some form of marketing can reasonably and successfully address.
How many times have you dug into a creative assignment to discover the underlying business issue is actually sales logistics, pricing or government relations? In other words, not a potent marketing opportunity just yet?
This earliest, murkiest stage of marketing is a productive environment for brand-side managers and agency planners to leverage tools like Bing and ChatGPT. Especially if the humans involved are inexperienced or new to the team, sector, or industry. Here, generative AI can play the role of tireless research assistant, inventive librarian or wise MBA instructor. Where the arrival of the Internet and then search removed tremendous barriers, AI-generative text unlocks even more.
The task is to distill. An MBA instructor might ask, “Have we identified a problem worth solving?” And you could, quite literally, use Bing or ChatGPT to pursue that line of inquiry.
PROMPT to identify a problem: “You are an MBA instructor. I am your student. You’re going to teach me how to research and write an actionable problem statement. The industry we’re focused in is X. The category is Y. The marketing scenario appears to be [Awareness, Loyalty, etc.]. In as few steps as possible, teach me how to distill and organize a marketing problem statement in the style of an MBA course. Use markup to make it easy to read.”
You might ask your generative tool to provide several different responses so you can collective perspective; then ask it (cut/paste) to contrast and distill a singular response.
You might compare and contrast responses to the same prompt from different generative tools.
You might try adjusting single words inside a prompt to see what changes.
Another approach you could try here at the beginning…
PROMPT to identify a problem: “You are an MBA instructor. I am your student. You’re going to teach me how to conduct a SWOT analysis. The industry we’re focused in is X. The category is Y. I want to understand how the top five competitors, defined by revenue, compare with each other. Use a chart for the SWOT analysis, and markup to make it easy to read.”
Again, mix and repeat across different platforms and with subtle variations in your prompting. I maintain an Evernote file which I copy/paste responses into as I go. Use whatever tools work best for you.
The habit here feels academic—using the tireless generative engine to hone a line of inquiry in a classroom. But you can do it at 3 a.m. and no one cares!
Once you’ve cleared out a working thesis from the wilderness, you can use the same tools to further refine your thinking and analysis.
PROMPT to clarify worthiness: “You are an [MBA instructor, private equity analyst, venture capital banker, etc.]. I am your [student, client, prospect]. I am going to present you with an actionable problem statement and you are going to help me evaluate its worthiness. I need you to be skeptical and poke holes in my thesis, to help me improve the statement. Use markup. Here is the problem statement: [COPY/PASTE].”
Edit, and regenerate as needed to continue the conversation.
Then eventually you might try…
PROMPT to frame solutions: “You are a [CMO, marketing instructor, etc.] and I am your [customer, student, etc.]. We have already clarified an actionable problem statement for a marketing program, and now you are going to help me discern ways to begin framing solutions to the problem. I need [number] of directions to frame solutions. Give me a headline for each direction with three supporting bullet points. Use markup. Here is the problem statement: [COPY/PASTE].” 👈🏽 And you’d use the refined statement from your worthiness prompting.
So that’s just one path you could take.
You might take apart various elements of business school pedagogy and use them within your prompting. For example:
PROMPT: “You and I run a company. We are trying to understand the primary drivers of the return on equity for our business. I need you to teach me how to conduct a DuPont analysis. [And you might include some clues about industry, sector, etc.]”
PROMPT: “You are a [CMO, MBA instructor, CEO] in the [pick one] industry. I need you to teach me how to conduct a marketing investment analysis so we can allocate media investment with greater efficiency.”
PROMPT: “You are a CFO. I am a banking analyst. You are trying to persuade me to invest in your company. First, you are going to diagnose the current liquidity of the companies in [pick one] industry. Please quote other CFOs from public earnings statements to validate your diagnosis.” 👈🏽 And you’d use Bing or Bard since they’re connected to the live Internet.
The point is to realize you are now incredibly enabled, and augmented. Your team can leverage its creativity to synthesize and discern useful insights to frame actionable business issues.
Onward!