How local is your global ecommerce strategy?
Instacart & OpenAI Will Transform Shopping. This is our Breakdown.
Instacart’s recently released ChatGPT plug-in, which as of May 31, 2023 is being rolled into the native search of Instacart, has massive implications for brands operating in the world of AI-enabled shopping.
What exactly is happening because of ChatGPT?
In the video above, there are a few key things you should notice that will fundamentally change how you think of the digital shelf, retail media, and overall shopping behaviors.
ChatGPT not only proposes a shopping list to the consumer and has it added to cart. It makes some key purchase decisions for them as well.
First, it decided where to shop — you’ll notice they added Publix Super Markets onion powder and club soda to the cart.
Second, it decided what brands to buy — ChatGPT selected Unilever Hellman’s Mayo, Ole Mexico La Banderita Tortillas, and B&G Foods Crisco Vegetable Oil.
This effectively eliminates the famous term FMOT or First Moment of Truth coined by A.G. Lafley (CEO of Procter & Gamble) which was described by P&G as the “moment a consumer chooses a product over the other competitors offerings.”
AI will now make that decision for you.
What are the current limitations of Instacart’s OpenAI application?
We performed a test to see how deeply integrated the ChatGPT app is into the Instacart database and found that as of the writing of this post (June 4, 2023), the app still relied on ChatGPT’s database of information from 2021 and did not take into account prices or discounts within Instacart.
This was our prompt:
Hey Instacart, I have $25 available and need to cook a healthy meal for 2 adults and 1 child tonight. One of us is gluten intolerant and the other doesn’t like chicken. Based on what’s on sale right now can you give me a recipe of something I can cook to fit within this budget and dietary constraints?
These were the results (you can see the transcript of our chat here)
We requested a meal that fits within $25, and it returned a recipe and shopping list for us.
Upon adding it to cart, the meal exceeded the $25 budget, so we went back asking for something cheaper. It returned a new recipe but it was still too expensive once we added the ingredients to cart (~$29 before tax and fees).
Note that we did this test in Los Angeles, California, where prices may be higher compared to other markets. We also picked Gelson’s which was the first retailer in the list; unlike the demo, it didn’t automatically pick a retailer for us.
Both of these could be reasons for why the proposal didn’t fit the budget we had given.
These problems are easily solvable, and likely with the native search integration launched on May 31, 2023, our hunch is that you’ll see pricing and sales incorporated into future iterations of this.
What questions should you be asking yourself going forward?
It’s only a matter of time before every single retailer implements some version of this into their purchasing experience. Here are a few questions you should be thinking about in your future strategy sessions:
- How should I begin to design my PDPs to “rank higher” on AI-recommendations?
- How should I better allocate my retail media dollars to impact how my products show up in this new AI-driven purchasing experience?
- What can I do in my brand marketing to influence prompts written by consumers?
What action can you take today to get ahead?
The rate by which generative AI has been showcasing innovation is moving at a breakneck pace. While it may be difficult to justify devoting resources towards something that seemingly affects only a small percentage of your sales, it is important to realize that if this does become the future norm of shopping behavior, the early investments you make today will pay dividends in both your team’s readiness and internal capabilities to thrive in this future world.
- Be an early adopter: Many retailers likely have already or will soon be rolling out alpha and beta programs for brands to partner and co-build these capabilities. Get into as many of them as possible to learn the limitations of these capabilities, as well as capture early wins as the pricing and advertising models likely are not fully optimized.
- Use AI to your advantage: CarMax used OpenAI to affect the ZMOT (or Zero Moment of Truth) when a consumer is researching a product before walking into store. They specifically used it to mine reviews to improve the equivalent of their PDPs, saving 11 years worth of work to add high quality content to all their pages and boosting SEO as a result.
- Build a culture of acceptance, not fear: The chances of artificial intelligence and OpenAI-like models not becoming the norm is incredibly low. The consumer experience is simply easier as these models operate using natural language, just like talking to a friend or family member. This is not a fad. It will become the norm, and you have to shift your and your team’s mindset to accept and leverage these tools for your improvement and success.
Instacart Rolls out AI-enabled Search
Update: Sept 7, 2023
Instacart’s latest update introduces AI-driven responses in search, which may present a short-term arbitrage opportunity for you.
Rather than solely searching for specific products or brands, shoppers can now inquire about broader topics like “Healthy Breakfast Ideas.” This innovation has brought about a significant overhaul in the way search results are presented, with products now neatly categorized into sections such as “Oatmeal” and “Smoothie Bowl.”
However, the search results still aren’t flawless. Jello-O pudding can be listed as “chia seed pudding,” crackers shows up under “avocado toast,” and even Jimmy Dean’s sandwiches, which most shoppers wouldn’t classify as “healthy,” appear under “other healthy breakfast ideas.”
And therein lies the opportunity.
At present, it remains somewhat unclear how specific brands rank within AI-enabled search. For those looking to expand their organic share of voice on Instacart, a new threefold challenge has emerged:
(1) at the retailer level,
(2) at the keyword level, and now
(3) also at the AI-defined category level.
Note that this feature hasn’t been rolled out to everyone yet, so it’s advisable to balance your efforts, as search volume for this new feature may still be relatively low.
Nevertheless, due to the uncertainty surrounding its mechanics, it could potentially yield short-term benefits for brands that can decipher its inner workings.
Keep a close eye and see how this development might impact your top-selling SKUs.