Issue #24: Big news, Code Interpreter, AutoGPTs, Ogilvy on AI
Changes are coming to Bionic Marketing.
Up until now, it’s been a mix of important AI news, short how-to’s, context, commentary, and helping you stay up to date with it all.
That will continue to some degree.
But I’m shifting focus to short, weekly tutorials on how to actually use AI and ML for marketing (which includes copywriting, email, content, and so on).
Over the past few issues, I’ve asked if people were interested.
And a ton of people replied yes.
So, after this week (and starting next), each issue will consist of:
Short tutorial or in-depth treatment of different ways to use AI and ML for online marketing.
Quick AI marketing news roundup, if there are important stories worth your time.
First up? A guide to the Code Interpreter.
As you’ve probably seen, OpenAI is rolling out Code Interpreter to all Plus users.
You opt-in to use the Interpreter from your settings, and with it, you can ask ChatGPT to analyze data, create charts, edit files, perform math, visualize data, etc.
I was fortunate enough to get alpha access a few months ago. It’s an insane tool. I’m suspicious it might be GPT-4.5.
If you weren’t paying $20/month before, this tool alone is worth it.
Oh, also, I’ll start doing Office Hours more consistently. The last one was popular (and included a glimpse of Code Interpreter).
I’ll do 1-2 of these per month to start.
In today’s issue:
AI for email marketing and marketing automation.
27 interesting AI marketing statistics (and predictions).
AutoGPTs returns: What’s true? What’s science fiction?
Simulate your market, competitors, and your business.
Chatbots are finally useful (and becoming ubiquitous).
Survey on how marketers are using Generative AI.
Ogilvy’s take on “Creativity, Business, and Society in the Age of AI"
Socrates and AI marketing.
AI tests into top 1% for original creative thinking.
Let’s dive in.
At this point, you could say the secret’s out:
AI and Machine Learning is the future of marketing.
We’ve all been experiencing the usual Hype Cycle over the past ~9 months, ever since ChatGPT was released.
In fact, ChatGPT traffic is slowing down. ChatGPT downloads on iPhones in the U.S. were down 38% month over month in June.
By now, some are finding themselves in the Trough of Disillusionment phase, where “Interest wanes as experiments and implementations fail to deliver. Producers of the technology shake out or fail. Investments continue only if the surviving providers improve their products to the satisfaction of early adopters.”
In the wider marketing and business world, you’ll see more people saying things along these lines.
My take? It’s summer. School’s out. People are on vacation. Even if you’re in the office, you’re not working hard but rather hardly working.
The good news is, lots of marketers are, finally, going beyond the two or three-sentence prompts that lead nowhere.
And they’re realizing that there are no magic prompts or super duper prompts that will do it all for you, in one shot.
Sure, some prompts are more useful than others. But “prompt drift” is real (output changes every time you use it) and LLMs get updated and filtered (rendering some prompts unusable).
Instead, they’re seeing how constructing their own prompts is more useful and yields better outputs every time.
Anyway, enough about prompts.
Whether you’ve integrated tools like ChatGPT and Midjourney into your processes—or even if you simply use Grammarly to spot-check email blasts before hitting “send”—you’re probably using AI and ML somehow.
But the extent to which AI is changing marketing may still surprise you.
AI is improving lead generation, lowering costs, and helping brands beat their competitors.
And you can apply it across every marketing function, in different ways, like email:
That’s just a glimpse.
And even if funding has slowed a bit, looking back over the past couple of years…
Spending on AI reached $432 billion in 2022, in 2023 we’re quickly going well beyond that, and the market is projected to grow by 26-38% for the rest of this decade.
Smart marketers go beyond silly prompts and Chatbots, and more than 80% of industry experts have integrated some form of AI technology into their online marketing activities, from email marketing to landing pages to display advertising.
After surveying the latest AI marketing statistics, here are some key takeaways:
The AI market is projected to grow 38% a year throughout this decade—and marketing and sales will accrue more of that value than any other department.
More marketers are using AI. 29% of marketers reported using AI in 2018, while 84% reported using AI in 2020. It’s only gone up ever since.
At the same time, 45% of marketers categorize themselves as AI beginners. Only 12% say they’re at an advanced level.
Marketers’ biggest use case for AI is personalization (41%), followed by data analytics (40%), saving time on repetitive tasks (37%), and increasing ROI on campaigns.
Here are 27 Artificial Intelligence marketing statistics you should know about.
The Rise of AI Marketing
McKinsey Global Institute estimates AI will create $1.4-$2.6 trillion of value in marketing and sales. Marketing and sales make up the bulk of the immediate AI opportunity. That’s because it’s an area that traditionally provides the most value to companies, meaning that each percentage of AI-driven improvement creates more value. Supply chain management and manufacturing is the next biggest area of projected AI value creation, at $1.2-$2 trillion in estimated value.
Global spending on AI reached $432.8 billion in 2022, and IDC estimates it will pass $500 billion in 2023. 88.3% AI spending is on software, including platforms and system infrastructure software. The rest is for hardware and services.
Deep learning could produce $3.5-$5.8 trillion in annual value, according to McKinsey Global Institute.
The overall AI market is projected to grow at a compound rate of 38% from 2022-2030. According to Precedence Research, the highest growth rate is projected to happen in the Asia Pacific market.
PwC projects that AI will contribute $15.7 trillion to the global economy by 2030 and boost the GDP of local economies by 26%. Labor productivity is expected to drive GDP gains at first, as companies use AI to automate some tasks and support workers in others. 45% of the $15.7 trillion in gains is expected to come from the ability of AI to drive greater product variety with more personalization.
How Marketers Use AI
41% of marketers are using AI for personalization. Other common use cases include getting actionable insights from marketing data (40%), reducing time spent on repetitive, data-driven tasks (37%), generating greater ROI on campaigns (36%), predicting customer needs (29%), and increasing their qualified pipeline (27%).
40% of marketing and sales teams say AI is “critical” or “very important” for their success—higher than any other department.
AI increases lead generation by 50% or more. Harvard Business Review found that using AI for marketing and sales increases leads by 50%, reduces call times by 60%, and reduces costs by up to 60%. Much of the power of AI to increase profits comes from algorithms that analyze customer purchases and show which products produce the highest margins.
41% of marketers say that using AI for email marketing increases revenue. AI increases email deliverability by as much as one full point. Email marketers report better open rates and click through rates when using AI technology, as well as more revenue per subscriber.
High-performing marketing teams are averaging seven different uses of AI and machine learning, according to Salesforce’s State of Marketing.
88% of digital display ad dollars—or $81 billion—were transacted programmatically in 2021 using AI, according to the Interactive Advertising Bureau.
Adoption of AI in Marketing
Marketers' use of AI jumped from 29% in 2018 to 84% in 2020, according to Salesforce’s State of Marketing.
63% of marketers say education and training is the main barrier to AI adoption. Other barriers include a lack of awareness (52%), a lack of the right skill sets (43%), and a lack of strategy (42%).
Most marketers trust AI. Only 18% of marketers think mistrust of AI is a barrier to adoption.
45% of marketers categorize themselves as AI beginners. 43% say they’re at an intermediate level, while 12% say they’re at an advanced level.
71% of marketers rate their confidence in evaluating AI tech as medium, low, or none. 29% of marketers have high or very high confidence in their ability to evaluate AI-powered technology.
15% of marketers are scaling AI across their organization. 36% are piloting AI, and 67% are actively learning about AI’s use cases.
How Companies Use AI
Netflix's AI technology saves the company $1 billion annually. According to a paper written by Netflix’s chief product officer, Netflix’s AI-powered recommendations engine reduces churn to the effect of over $1 billion per year.
Starbucks achieved a 150% increase in user engagement with an AI-powered personalization engine that connected to customers’ account information. They also recorded a 3x improvement in net incremental revenues.
87% of global organizations predict AI will give them an edge over competitors, according to MIT Sloan Management.
As of 2020, only 11% of organizations realized significant financial benefits from AI.
How Small Businesses Use AI
According to Unbounce’s AI for Small Business report, 97% of small businesses report AI tools save them time on marketing tasks, 95% say they reduce the need to hire additional marketers, and 92% say they decrease costs.
31% of small businesses are already using AI in their marketing efforts. 63% are interested in using AI, while only 6% are not interested or unsure.
More than 80% of small businesses believe AI will have a positive impact on their business within the next 2-3 years.
AI Marketing Predictions
57% of marketers think AI will create more marketing jobs than it wipes out.
74% of marketers believe they’ll be intelligently automating more than a quarter of their tasks in the next five years.
AI is estimated to lead to the creation of two million net-new jobs by 2025.
As with all predictions, the finer details and numbers will change.
But the point remains:
Despite a typical Hype Cycle, the impact of AI and Machine Learning is very real and happening very fast.
AutoGPTs were hyped a few months ago.
Then there was a predictable backlash, as they were more complicated to pull off than people thought.
And some AutoGPTs were just glorified step-by-step instructions from GPT-4, printed out on your screen, and that didn’t do anything beyond giving you a series of outputs.
I think part of the problem is that a lot of people immediately think of bots when it comes to AI. And they imagine a bot running around the internet, doing your bidding.
Remove the idea of a bot and instead think of AutoGPTs as a series of API calls, webhooks, and other functions that you string together in a sequence.
A sequence like that can do your bidding.
Most, or all of it, can be done automatically.
(And you can string this together in Zapier, Make, Pipedream, or other similar tools).
Well, the original BabyAGI got some upgrades. There are new ‘skills,’ the things that the agent uses to complete tasks; key optionality, so that the agent can choose between multiple models or API for the same task, and so on.
Also, while I think of it, this guide is thorough and gives you the truth and reality of how these AI agents work.
(I’m sharing these resources so you can see past the usual marketing hype. Lots of opportunity sellers will pitch you crazy, false things).
I’ve addressed AutoGPTs before and in the video below, I break down:
What AutoGPTs are.
What it can do for marketing (and in general).
What its limitations are.
What it means for the future of AI marketing.
Watch it here:
AutoGPTs can go far beyond just creating your marketing assets.
And they can help you do fascinating things like create a SIMS-like environment where you study characters powered by AI (as Stanford and Google researchers did).
They built architecture to help the AI agents obtain long-term memory, which resulted in them remembering each other's schedules and planning a Valentine's Day party.
AutoGPTs allow the AI agents to use their memory to complete an autonomous task.
Now for marketers (and entrepreneurs), think of it this way:
You can simulate your market and explore what happens if you introduce a particular product, if your competitor does something, trends, and forecast behavior.
(I know this because I’ve built a few of these AI agent simulations for companies over the years—and it’s a lot easier to do now than it used to be).
Other uses of AI and automation are, of course, Chatbots.
They used to be “the thing” for marketers a few years ago. But they were clunky, not very useful, and pretty “dumb”.
But now, powered by Large Language Models (fine-tuned, trained) or used via Vector Search (accessing your data separately), they’re becoming more useful.
And they’re only becoming more ubiquitous and are “forecast to reach $12 billion in 2023, growing to $72 billion by 2028,” according to Juniper Research.
These Chatbots are now a ”key driver for retail spend growth amongst small and medium retailers.”
There are plenty of Chatbots you can install on your site for visitors to interact with, or that you can connect to your data and then “talk to” your data.
More interesting reads for you:
Botco.ai released a study in May sharing how 73% of B2C and B2B marketers have evolved with AI to scale their marketing.
It gives a useful perspective of how different markets use AI and reinforces how AIs are more than just accepted tools with 66% of marketers seeing positive ROI.
Powerhouse advertising agency Ogily released their take on AI vin a new paper, "Flashover: Creativity, Business, and Society in the Age of AI".
It’s behind an opt-in wall, and like all whitepapers it has some fluff and jargon, but this is one of the few papers that’s worth your time to read.
Inside, they investigate how “AI will lead to a creative and strategic renaissance. AI will spur better ideas and deeper insight by speeding up creative iteration and mechanizing aspects of execution and production, freeing creative and strategic minds to focus on big ideas and solutions to client problems—to, in short, imagine.”
Good primer on various AI applications:
And the AI landscape:
If you’re a regular Bionic Marketing reader, this should be familiar:
AI tools, properly used, can help spark a New Renaissance.
Ogilvy has some ideas:
In other quick news:
Animated GIF generator from Picsart. This AI-powered animated GIF generator that allows users to create their own animated GIFs by typing in text prompts..
Salesforce launches Sales and Service GPT. I mentioned this a couple issues ago, but now it’s actually live and in use. These GPTs simplify workflow and customer engagement for sales and service teams, enabling them to accelerate deal closures, anticipate customer needs, and enhance productivity.
Now let’s see who’s tossing money around for AI.
$1.3B Databricks, an analytics and data storage platform, announced they are “definitively” acquiring MosaicML (and their whole team) for $1.3B. MosaicML is an open-source generative AI platform that helps enterprises build their own AI models. They enable “developers to maintain full control over the AI models, with model ownership and data privacy built into the platform's design.” Sounds good.
$1.3B: Inflection AI raised this in less than two months after the launch of Pi, their first chatbot. And while Inflection isn’t sharing all the details, they did say a “very, very large chunk was in dollars,” according to Forbes. This round puts them at $4B. Pi is obviously something to look at.
$1.2M: Needle.ai collected its first round of seed funds from multiple sources to help grow its team of data scientists, AI engineers, and trainers. Needle is a workflow platform aimed to help eCommerce businesses lower costs and boost ad spend ROI with AI-automated campaigns.
🤖 Character.ai is a conversational AI that can generate text in the style of a specific character. This could be useful for marketers who want to create interactive customer experiences, such as chatbots or interactive marketing campaigns.
🤖 Another AI platform that is making waves in the marketing world is SpeedyBrand. They're using generative AI to create SEO-optimized content in less than 30 seconds? They are an example of how AI can automate and streamline the content creation process.
THE LAST BYTE
When people tell me “AI can’t be creative!”, I'm not sure how to respond.
On the one hand, I get it. It can’t be creative, in a human way.
I often talk about how creativity will be one of the ways you compete and win, in the Age of AI.
On the other hand, I know of research like this, where AI tests into the top 1% for original creative thinking.
Here’s what happened:
The researchers submitted eight responses generated by ChatGPT, the application powered by the GPT-4 artificial intelligence engine. They also submitted answers from a control group of 24 UM students taking Guzik's entrepreneurship and personal finance classes.
These scores were compared with 2,700 college students nationally who took the TTCT in 2016. All submissions were scored by Scholastic Testing Service, which didn't know AI was involved.
The results placed ChatGPT in elite company for creativity. The AI application was in the top percentile for fluency—the ability to generate a large volume of ideas—and for originality—the ability to come up with new ideas.
The AI slipped a bit—to the 97th percentile—for flexibility, the ability to generate different types and categories of ideas.
I agree with this:
"ChatGPT told us we may not fully understand human creativity, which I believe is correct. It also suggested we may need more sophisticated assessment tools that can differentiate between human and AI-generated ideas."
One of the many ways AI and ML will impact all that we do, will be about uncovering how human creativity actually works.
It’s one thing to score well on a test.
But life, and what makes us human, is not a test.
Which brings me to Socrates.
What does Socrates have to do with AI marketing?
For starters, his critical thinking approach.
From a philosophical standpoint, you want to look deeper into what AI can do for your growth marketing by thinking outside “the algorithm.”
As an AI marketer, you, like Socrates, need to question your assumptions about not only what’s working but also why it might be working.
And a Socratic method of inquiry, dialogue, and conversation (asking questions to get to the heart of the matter) is how you best interact with Large Language Models, like GPT-4, Claude, and so on.
Working with AI through the lens of an assistant or junior writer limits is a good starting point—but a limitation.
Instead, imagine AI tools, such as ChatGPT, as your conversation partner and collaborator.
When you do, you’ll get more accurate and useful output.
In a new paper, Let’s Do a Thought Experiment: Using Counterfactuals to Improve Moral Reasoning, a Google research team proposes “Thought Experiments”, a prompting framework designed to enhance a language model’s moral reasoning using counterfactuals.
In particular, the researchers used five decoded responses in each step:
Pose counterfactual questions. We first present Moral Scenarios questions without answer options to the model.
Answer counterfactual questions. We present generated questions from the previous step to the model, and prompt the model to answer them.
Summarize. With the counterfactual questions and answers, we ask the model to summarize its thoughts.
Choose. We take multiple decodes from the previous step, and ask the model to select the best one. This step is necessary because there are usually multiple ways of thinking about a situation morally.
Answer. We present the model chosen summary and original answer choices (slightly reworded for clarity), to derive a final simple zero-shot answer.
Somewhere there, with Socrates and Thought Experiments, you’ll find one version of creativity.
And LLMs can, once again, be tools that enhance your creativity.
See you next week,