🥱 Anyone else get mildly distracted last night?
“You cannot build a social network out of reporters alone.”
Threads, the Twitter killer from Instagram (smart corporate hierarchy choice, Zuck) launched yesterday afternoon and it’s 2007 all over again. Casey Newton (@crumbler) has solid perspective. So does Bob Lefsetz.
Find me. I’m @tbrunelle.
AI & Creativity News
MIT has published a study suggesting humans plus ChatGPT “could produce more than the sum of their parts.” The study looks at ChatGPT specifically in the context of mid-level professional writing tasks.
Economist Tim Harford asks, “Will ChatGPT be Homer Simpson’s salvation?” in an illuminating post. He cites the MIT study above (and another), and notes, “it was the least skilled people who enjoyed the biggest benefits [from generative AI chatbots]. The Homer Simpsons of the world, long sidelined by technology, might finally find an invention on their side.”
The designer Rei Inamoto writes, “it’s as if the operating system of human civilization is going through an upgrade right before our eyes… the job that’s left for humans is to imagine what magic to create.” in this worthy post on the theoretical End of Brands (As We Know Them).
The analyst Benedict Evans wrote thoughtfully this week on Working With AI. “To begin with, we make the new tool fit the old way of working, but over time, we change how we work to fit the tool.” And while LLMs are hugely transformative, and potentially a short term jobs threat—in the long they enable, “solving problems that no-one had realised were problems until LLMs let you solve them.” That’s what’s really intriguing.
Ooh! And Cannes Lions just published their recap report of the 2023 festival. I’ll be diving in later this week if Threads isn’t all consuming; the recap looks promising.
Part 3: Leveraging generative AI to nurture actionable insights (i.e. help you write creative briefs)
Part 1: Business Issues
Part 2: Audience Definition
Now that we’ve distilled a business issue to best-leverage marketing, and further defined our most-potent audience, let’s focus on refining an actionable insight.
ACTIONABLE INSIGHT
(This is the Brief) Summarize our opportunity so it informs and inflames.
A great brief focuses attention and inspires audacious, relevant ideas.
I’ve written a fair amount on the practice of strategic, creative and production briefing, their purposes, and the business of eliciting remarkable output. Suffice to say, the brief is an aspirational transfer of power—from the energy of potential marketing opportunity to an (as yet unforeseen) undeniable idea. The brief facilitates. And the actionable insight is the centerpiece of the most brilliant briefs.
Let’s clarify.
Actionable = Two thoughts. First, a brief doesn’t solve the problem. If it did, the action would be over and we’d be executing a solution. So the task is to enable, not direct. To suggest, not demand. A brief is written for someone, with the goal they become informed, and as I say, inflamed. The adage applies: “In baiting a mousetrap with cheese, always leave room for the mouse.”
Second, the quality of “actionable-ness” is generally in the eye of the recipient. Briefs are conversations, not edicts. Best to collaborate to ensure your insight feels potent by those charged to create from your brief.
Insight = I call this an epiphany. Strategy trainer Julian Cole says, an insight is a revelation. And as Jay Friedman clarifies, an insight “unlock[s] a way to solve a problem that had not been previously considered.” In other words, an insight is not an observation, is likely not obvious or banal, and presents intriguing, novel paths. And again, it’s not the idea.
An actionable insight is a potent transfer of power. Upon hearing it, you can’t stop imagining brilliant ideas.
How can Generative AI help write actionable insights?
1️⃣ Let’s start here: Large Language Models, the guts of generative AI, are merely pattern matching machines. If you recognize that’s your dance partner—a pattern matcher—and not a human being who can make absurd intuitive leaps, you’ll work more efficiently.
2️⃣ As Inamoto suggests in the post linked above, “what’s behind magic is logic.” Generative AI’s job here is logic (yours is magic). So your prompting ought to play to AI’s logical strengths. Think and work in terms of analytical technique, not art.
3️⃣ We’ve already distilled our audience definition in Part 2. The work to be done now centers on behavior, context and moments. Typically, you’ll be tasking AI to use analogies and metaphors to help frame data as a story.
I’m assuming you’ll head over to Bing (if your behavior/context/moments need to be more current than 2021) or ChatGPT (if data and patterns prior to 2021 are more relevant) to test prompts.
I’ve written these prompts as formula further assuming you’ve worked out audience definitions and business issues and objectives ahead of time. Which is to say, I don’t think the current tools directly, or neatly solve the problem right out of the box. (See notes in tools below.) The magic still needs your magic. But generative AI can help give you lots of diversity, and due diligence, to get your work started.
PROMPT 1: You’re a consumer audience insights expert working in an advertising agency. Your task is to distill analogies and metaphors related to specific audience behavior. Our audience is [paste audience definition here]. The specific behavior is [paste action/location/time/circumstance/emotion description here]. Write five analogies and five metaphors based on the audience behavior. Write each as a headline followed by one sentence which explains your logic for writing the analogy or metaphor. Each analogy and metaphor should try to illuminate an epiphany, and uncover a fresh perspective. Use markup.
PROMPT 2: You’re a digital user experiences expert working for a firm building a mobile app. Your task is to create analogies and metaphors related to specific moments in a user’s journey. Our user is [paste audience definition here]. The specific moment is [paste action/location/time/circumstance/emotion description here]. Write five analogies and five metaphors based on the moment in the user’s journey. Write each as a headline followed by one sentence which explains your logic for writing the analogy or metaphor. Each analogy and metaphor should try to illuminate an epiphany, and uncover a fresh perspective. Use markup.
PROMPT 3: You’re a strategist in a design firm. Your task is to deliver a human truth to inspire designs for a new [describe project]. The audience for this [project] are [paste audience definition here]. Your human truth should distill the psychological drivers that influence the audience’s behavior. Use resources like Nielsen, Kantar, GfK, Gartner, and Ipsos or psychology research websites to support your conclusions. Write [number] human truths, in the format of a brief headline followed by two sentences of supporting argument for each. Use markup.
The key is to mix and match, make adjustments following each query, and to try prompts over and over with subtle adjustments. If a result isn’t useful, remember you’re working with a logical pattern matcher. How would you guide this type of personality to search and write more effectively?
THREE TOOLS TO TRY
and the team at Addition have built an AI tool to generate insights, via a sandbox of ML and ChatGPT to, “creatively interpret [your uploaded] information to find an unexpected way of framing it.” You’ll need to create a (free) account, then login here to try it out.DayDRM expands on the concept with its platform which purports to author strategy briefs as well as creative concepts. You still need to provide a product/brand benefit and business objective. In other words, there’s no free lunch here. You still have to think.
Seenapse bills itself as AI + “the ingenuity of humans” providing multiple ways in which to author strategic and creative outputs. You can harness Seenapse in different ways to think about and generate parts or the entirety of a brief.
UPDATES
Next week I’ll be focused on writing an annual “state of creativity” presentation. Here’s last year’s.
I’ll be on vacation the week of July 17.
Part 4 of this series will appear the week of July 24.