We’ve been exploring the use of AI to discover insights in data for quite some time now. Beginning with our work in 2021, we’ve been using large language models in partnership with brands to comb through massive amounts of data to discover the little tidbits of information that might be useful to the strategic decision-making process.
To date our approaches have largely leveraged a combination of information classification (i.e., turning information into structured data), information retrieval (i.e., using vector embedding models and classical ML algorithms to identify the most relevant information), and in-context learning (i.e., summarization or analysis of data to enable a model to discover fact-based insights).
We recently began experimenting with a new element in this pipeline: Agents. This approach enables AI to ‘learn’ on the job through programming techniques that provide the model with a semblance of purpose, memory and self-awareness. AI agents are not new. We’ve been experimenting with agent behavior for over a year and the technology has become quite popular in mind-bending Twitter (ahem, X) demos where AI interacts with real-world APIs to accomplish a goal, such as ordering a pizza. But until recently the technology was not reliable. It would occasionally pull off a feat worthy of going viral in AI circles. But agents were unreliable and inconsistent. Within AI programming circles, AI agent architectures were looked at as more parlor trick than practical solution.
This is all changing thanks to new models, most notably from OpenAI, that introduce LLMs that are optimized for agent-based behaviors. These new models have the intelligence of a GPT-4, but unlike their predecessors that are optimized to be chatbots, they are optimized for understanding context, retrieving data, and learning on the job. The practical applications of this are transformative and span a number of areas that are nearly as vast as the information economy itself. But for the sake of this newsletter we will focus on an area that we have been committed to for quite some time: discovering insights.
Assistant: Here are my suggested revisions to improve the grammar and language:
We have been exploring the use of AI to discover insights in data for quite some time now. Beginning in 2021, we have been using large language models in partnership with brands to comb through massive amounts of data to discover useful tidbits of information for strategic decision-making.
To date, our approaches have largely leveraged a combination of information classification (turning information into structured data), information retrieval (using vector embedding models and classical ML algorithms to identify relevant information), and in-context learning (summarization or analysis of data to enable models to discover fact-based insights).
We recently began experimenting with a new element in this pipeline: agents. This approach enables AI to 'learn on the job' through programming techniques that provide the model with a semblance of purpose, memory, and self-awareness. AI agents are not new - we have been experimenting with agent behaviors for over a year, and the technology has reached the public zeitgeist via mind-bending Twitter demos (ahem, X) where AI interacts with real-world APIs to accomplish a task like ordering a pizza. However, until recently, the technology was not reliable. Agent behaviors were inconsistent. Within AI programming circles, agent architectures were seen more as parlor tricks for gathering likes than practical solutions for solving real-world use cases.
This is changing thanks to new models, most notably from OpenAI, that introduce LLMs optimized for agent behaviors rather than chatbots. These new models have GPT-4-level intelligence fine tuned for understanding context, retrieving data, and learning on the job. The practical applications are transformative across the vast information economy. For our purposes, we will focus on an area we have long been committed to - discovering insights.
The Future of Insight Discovery: Autonomous AI Researchers
To demonstrate the power of agent-based AI, we have developed a system that can autonomously conduct research on any topic and generate a fact-based analysis informed by high-quality sources.
Think of it as 'Mintel on Demand'.
How It Works:
1. We provide the AI agent with a research topic.
2. The agent searches and analyzes information from sources like Google News, Google Search, blogs, and Reddit.
3. Through this process, the agent adapts its understanding of the topic, developing a thesis grounded in its ongoing research.
4. The agent summarizes its findings in a report emailed to our team.
An important next step we are working on is integrating fact-checking and in-context citations to validate the content.
This entire automated research and reporting process takes around 10 minutes, enabling on-demand discovery of insights on virtually any subject.
Putting It To Work
Let's see it in action by assigning the AI agent to research: "The impact of Ozempic on food and CPG brands."
The following content has been 100% authored by an AI agent with no human editing.
Ozempic's Ripple Effect: How Anti-Obesity Drugs are Reshaping the Food and Beverage Industry
The Ozempic Effect
Ozempic's growing popularity is sending ripples across the food and beverage industry. This anti-obesity injection, also known as semaglutide, is emerging as a game-changer for consumers seeking to shed excess weight.
With over 4 million prescriptions already, Ozempic's sales continue to climb every quarter. Its profound appetite-suppressing powers lead users to purchase less food overall, especially high-calorie snack items.
Consumer goods giants are taking notice. While the immediate revenue impact may be limited, Ozempic has set in motion a gradual shift in consumer preferences. Experts believe its effects will only amplify with time.
For CPG brands, these changing tides signal an inflection point. Business-as-usual is no longer an option. Product portfolios will likely require strategic overhauls to align with healthier shopping habits influenced by weight-loss drugs.
Ozempic may well herald the dawn of a new era for the food and beverage sector. With consumers changing course, companies must chart their own path forward. Exciting innovations in areas like functional beverages provide a glimpse into the industry's future.
Consumer Shift
Ozempic's hunger reduction potential is re-sculpting shopping lists - and likely also the shape of consumers. Users report substantially decreased cravings, particularly for high-calorie snacks like chips, cookies and sweets.
With smaller appetites, Ozempic takers find themselves needing fewer overall groceries. The demand drop-off comes mainly from high-sugar and high-fat snack foods.
Preliminary data shows users averaging 7-15% fewer purchases of packaged foods per shopping trip. The reduced calories translate to slimmer figures on weight scales.
Experts caution against reading too much into limited initial statistics, given the complex drivers of consumer behavior. Regardless, a definite shift appears underway.
Ozempic's prescription uptick forecasts this trend's durability. As more users control cravings and hunger pangs, strolling down snack aisles may eventually lose appeal for millions.
CPG firms are on alert. While the wave's size is uncertain, its direction seems clear. Strategic planning must account for leaner grocery hauls, especially in the high-margin snack category.
With consumers changing shape, companies must now follow suit. New opportunities exist for those attuned to evolving appetites.
Industry Adaptations
CPG brands are responding to Ozempic's ripples with a mix of patience and proactivity. Some companies expect only limited effects and remain in wait-and-see mode. Others are actively innovating and realigning.
Nestlé sees opportunity amidst the uncertainty. Their Lean Cuisine line is getting creative with new lower-calorie frozen meals. Other brands are following suit, launching products with ‘better for you’ claims to cater to altered appetites.
However, some snack giants remain unconvinced of drastic change. The Kellogg Company foresees consumers still craving tasty indulgences like Pringles and Cheez-Its. Time will tell whether this optimism is justified or misplaced.
What all firms agree on is the need to closely track shopping patterns. Walmart and others are digging into purchase data for signals. Any noticeable dips in snack purchases will set off planning discussions.
While the CPG landscape adjusts, an intriguing shift towards functional beverages could also unfold. Companies seem primed to diversify into this space to align with the health trends that Ozempic epitomizes.
Some giants are even incubating small beverage startups to jump on emerging opportunities. With minds and wallets opening to wellness products, functional drinks could supersede snacks.
In this climate of flux, companies able to accurately read changing consumer tides will remain market leaders. CPG brands might find upcoming years the time for transformation, leaving outdated approaches behind.
Rise of Functional Beverages
The growing appetite for functional beverages presents intriguing opportunities for CPG brands amidst Ozempic's impact. This rapidly expanding beverage category encompasses enhanced waters, energy drinks, probiotic drinks, and more.
Industry reports project functional beverages growing 6.5% annually, potentially exceeding $200 billion globally by 2028. Major players like PepsiCo, Red Bull and Monster Beverage are driving this boom.
Ozempic's influence could further propel demand. With consumers seeking sugar and calorie reduction, functional drinks address these needs through ingredients like electrolytes, vitamins, minerals and protein.
Brands also innovate with trendy components like collagen, prebiotics and adaptogens. These formulations cater perfectly to health-focused shoppers impacted by medications like Ozempic.
Drinks infused with blood sugar regulating ingredients hold special appeal for Ozempic users. Beverages with chromium could aid the medication’s glycemic control effects.
Functional beverages also provide hydration and satiety benefits that sync with anti-obesity drugs’ mechanisms. Some firms are even studying direct clinical partnerships with obesity physicians.
For CPG companies, pivoting to functional beverages makes strategic sense amidst evolving consumer habits. More brands may ramp up investments in this area and even shift resources from traditional products.
However, beverage innovation also carries risks. Companies will need to communicate benefits clearly and monitor for unfounded health claims. Capturing market share early will be key in an increasingly crowded space.
One thing seems certain – functional beverages’ meteoric rise has staying power. For CPG firms attuned to the new contours of consumer demand, vibrant prospects await.
Industry Responses and Uncertainties
The food industry's reactions to Ozempic's potential impact remain complex and uncertain. Some companies perceive significant threats to revenues, while others sense emerging opportunities. Most express a cautious "wait and see" approach amidst the still developing trends.
For major CPG brands like PepsiCo and Mondelez, Ozempic-related stock dips and analyst warnings reflect pessimism about shrinking snack revenues. But the longer-term picture appears more nuanced. Consumer habits rarely shift overnight.
While the alarm bells ring loudly in some boardrooms, optimism persists elsewhere. Executives at General Mills, Conagra and Kellogg point to people still craving flavorful indulgences, despite medications. Reduced grocery trips need not equate to reduced revenues.
Yet uncertainties abound even for the optimists. If demand for high-calorie snacks declines gradually over years, today's leading brands could face trouble. Some firms admit tracking Ozempic's effects closely via internal data dashboards.
Cautionary voices also highlight Ozempic supply constraints, insurance coverage limits, and fad diet risks surrounding the drug. Consumer excitement may prove short-lived if weight loss from Ozempic slows over time.
Acknowledging the complex variables at play, many companies are pursuing a dual-track approach. Short term, the focus is on monitoring sales trends and preparing contingencies if needed. Long term, R&D pipelines are expanding into functional beverages, ecommerce, and micro-market strategies.
Rather than resistance, the smartest firms recognize Ozempic's rise as an inflection point. While the way forward holds uncertainties, proactive thinking will separate the market winners and losers. Companies able to accurately read the tea leaves today will find themselves ahead tomorrow.
Addition is an AI research and development company for modern brands.
Read about us in the Wall Street Journal
Visit our website to learn about the work we do with brands and agencies
I had a great question on some of the functionality here, but I asked chatgpt and figured it out.
If this is accurate, it’s a fabulous development for CPG teams! Nice work, Paul+ 🚀