websights Skip to main content
Industry Insights

Food as Preventative Health?  

By January 23, 2024February 15th, 2024No Comments

Enabling Precision Nutrition through Pricing  

This series will feature posts from multiple Retail, Healthcare, and Technology thought leaders. Thank you for tuning in.

The Rise of AI-Based Shopping Agents, Part 1
By: Tim Ouimet, Engage3 Co-Founder

Welcome to our Vision for Personalized AI Shopping Agents! The vision lies at the behavioral intersection of: 

  • Shopper emotions 
  • Shopper literacy,  
  • Personalized commerce,  
  • Market transparency,  
  • Precision nutrition,  
  • Precision manufacturing, and  
  • Personalized health. 

It’s a vision where the shopper leads through everyday purchasing decisions. A vision where the cost structures, scale, strategy, and innovation of ecosystem stakeholders are predictably aligned with shopper values to provide outcomes that improve lives, communities and sustainability while also improving profit. Enabled by market transparency (data), Artificial Intelligence (knowledge) and outcome-based decisions (mathematical optimization), the potential for all participants (and mostly shoppers) to benefit measures into the trillions ($) and promises to reshape our retail ecosystems at scale by shifting the marketplace paradigm from products to people and from prices toward values. This shift will create a race-to-the-top and will come faster and be more transformative than expected. 

As Co-Founder of Engage3, I am pleased to kick this series off and am energized to be at the forefront of this power shift toward a shopper-led future. 


Brands are built on trust; a positive perception of what will be delivered. The limits of pricing power (sustainable profits) are found in the dopamine rich behavioral domain of emotional perceptions. 

Since the beginning of commerce, prices served as interpreters of value. Shoppers want to know the price and stakeholders want to know what a shopper will pay. As such, price intersects the decision-making fulcrum of all marketplace participants.  

We’ve all felt the thrill of discovering a price that emotionally compels a buying decision. Similarly, we’ve all felt the opposite. Discovering a price that compels a full stop. The emotional potency of a price can trigger a release of dopamine or disgust. The actions that follow often leave a behavioral trace in the archeological layer of historical sales data. 

Properly understood, the emotional potency of price is a dominant key to measuring the why behind a buy and to managing trust. From this perspective, the promise of AI is to leverage price as an interpreter of value that leads the marketplace focus from products to people (from prices toward values). 

The Promise of AI Agents


Shoppers want to be understood and served. Price signals value but value is more than price. With purchase behavior, shoppers often voice their values via a willingness to pay for service, quality, and convenience relative to the price of comparable competitive offerings. In retail, this is known as the Shopper Value Equation and is written as: value = service + quality + convenience – price. When quantified relative to the emotional potency of price, the non-price aspects of shopper value are understood with more precision. This understanding gives industry the knowledge to differentiate on value. 

Today’s personalized offer generation methods leverage vast amounts of data from various sources, including purchase histories, browsing histories, and social media interactions, to create detailed customer profiles. While data-rich, most methods focus on shopper segmentation. Some methods attempt to model price elasticity, but none personalize toward perceived value. These methods are challenged by a lack of transparency, a lack of knowledge and a lack of outcome-based decision-making. This lack leads to guesswork . Guessing is risky. In retail, it creates an over-reliance on price to close sales. To avoid upsetting shoppers, retailers align pricing with competition, not value. With this, retailers lack the resources to fully align organizational resources to serve shopper values.  This lack then leads to an undifferentiated race-to-the-bottom. 

Differentiating on non-price attributes with greater predictability and confidence leads a company to translate organizational strengths, strategies, and innovation into offerings that align with shopper values. In this way and properly interpreted, the shopper’s voice serves to align supply with demand in a mutual race-to-the-top. 


Shoppers are increasingly more attribute focused and literate. With this, product-attributes become the new “shelf space”; because that’s where attention goes (product-claims, nutrients, ingredients, packaging, company ethics, etc.). In the domain of attention, even digital attributes are limited. In this domain, shopper relevancy requires knowledge-based prioritization of the attributes or attribute bundles that will predictably drive a sale. 


Historically, human buy-side and sell-side agents play a pivotal role in commerce. Trusted buy-side agents serve as guiding lights for shoppers; helping them find, choose, and afford products that best meet their values while minimizing risk and saving shoppers both time and money.  

Trusted sell-side agents tune into shopper values to differentiate relative to comparable alternatives, to close sales while achieving organizational targets, and to structure customer feedback for better future alignment of the company’s products and offerings to demand. Yet (metaphorically), without a knowledge of shopper value perceptions relative to price and comparable alternatives, the shopper is talking but no one is listening. 

The scale of increasing complexity along the intersections above makes knowledge-based outcomes on both the buy and sell side humanly impossible.

Tipping Points are Behavioral


Shoppers now adopt new technology faster than retailers. While still a manual process, price-transparency empowers shoppers to compare pricing on demand. Expect B2C comparison shopping to progress toward an automated AI driven experience.  

In retail, Price transparency and informed shoppers already put significant downward pressure on retail margins. Today’s inflationary environments only increase people’s sensitivity to price.  

For a sense of the transformative potential of transparency, AI, and mathematical optimization in retail, consider the impact that High Frequency Trading (HFT) and behavioral momentum models had in the capital markets. In capital markets, investors measure market efficiency in terms of a Basis Point (BP) spread (where the BP spread is the difference between the bid price and the ask price). With HFT and behavioral models, the basis point spreads between the buy-price and ask-price collapsed by 30X and now trend at about 5 basis points, or less (equal to 0.05% or less). 

In retail, the average basis point spread between the high price and the low price for a given UPC is 6,100 or 61%. When comparing like-for-like brands, such as brand to store brands, the spread is even greater.  

With 9% price inflation, a typical retailer will see profit decay to zero in 1-3 months (without active price management). Simply raising prices does not account for how competitors will respond or predict behavioral changes in shopper switching behavior.  

With AI empowered shoppers, BP spreads in retail will tighten. While retail may never reach the efficiency of the capital markets, a one-order of magnitude change is possible and makes retail pricing practices ripe for disruption. Retailers need a behaviorally sensitive market response. 


Retail pricing offers a natural integration point for managed health care providers to motivate better FaM choices at scale. Many recognize the potential for FaM to materially reduce cost-of-care, acknowledging “you can’t medicate your way out of a bad diet.” Many also recognize the dominant role price plays in driving demand. 

Chronic diseases are on the rise, and current healthcare spending is not slowing this growth. In fact, US healthcare spending for chronic diseases that may be diet and lifestyle mediated amount to an estimated $3.1 trillion per year (72% of the nation’s $4.3 trillion annual healthcare costs). Meanwhile, Americans spend over $0.95 trillion annually on food and beverages. So, this means that for the estimated 60% of Americans with a chronic disease, there are $6.8 healthcare dollars spent for every dollar spent on food and beverages—and this ratio is growing fast. 

Both views above are strongly recognized[i]. While there are programs that demonstrate quantitative FaM outcomes, these programs are limited in scope and scale. Programs, such as: Geisinger Fresh Food Pharmacy, Produce Prescriptions and medically tailored meals all demonstrate positive patient outcomes with lifestyle diseases while lowering managed care provider liabilities. Even so, FaM has yet to gain broad adoption or scale. 

Today, FaM is about brute force techniques: billing codes, vouchers and a lot of effort and complexity for the shopper to become aware of and access an incentive (i.e., only about 1 in every 5 that are approved for SNAP use it). There are apps that support education with product scoring (i.e., NuVal, Diet ID, Guiding Stars, Tufts Food Compass, Yuka.io, etc.). Plus, there are said to be over 450,000 mobile apps in support of better health and food choices.  

FaM needs to become, but is not yet about, letting the shopper lead through simplicity. By simplicity, we mean:  

  • find (alerting a shopper to emotionally potent offers),  
  • choose (providing knowledge about why a recommended choice is superior),  
  • afford (making a superior choice price competitive with a less desirable preference) 
  • relevant (recommendations consider personal preferences), 
  • risk (de-risking trial of new product choices), 
  • adopt (the ability to download, configure and use an app), and  
  • use (an ability to shop all household products, not just approved UPCs). 

Further challenges include data integrity, standardized KPIs, validation of behavior change, and demonstrable cost-of-care outcomes. The lack of quantitative outcomes points to a need for a test and learn environment that can operate at scale and provide methodical and continuous control versus focus measurement relative to cost-of-care KPIs and outcomes. Thus, enabling stakeholders to take calculated risks to quickly adapt and learn and drive behavior change with guaranteed ROI. 

The areas above would all benefit from precision information, AI and optimization methodologies with closed loop machine learning and forecast to actual plus focus vs control reporting. The possibility is to link emerging nutrition science with product ingredients and nutrients for actionable personalized food-as-medicine recommendations. Then, to integrate insurer approval process into an offer generation engine; moving the approval process ahead of the retail POS system to provide a more actionable and immediate experience. A pay for performance model guarantees positive ROI. 

The trillion $ opportunity is to create a win-win-win by shifting inefficiencies that live within insurance company business models into retail transactions. The result: a win for the shopper, retailer, brand, and insurer. A pay for performance model guarantees positive ROI to the insurer. In this way, insurance companies provide funds and get quantifiable ROI on cost-of-care outcomes.  

[i] Managed care and retail are already aligning on this intersection with the Food as Medicine Summit and the NGA Nutrition Incentives POS Working Group (billing codes are coming).