See how Wallaroo.AI helps unlock AI at scale for Retail >

Why CPG Companies Need to Get AI And ML Right

Image courtesy of Jorge Royan

As a CIO or as a Corporate CTO, I was often pitched technical solutions in search of a business problem, or a solution to an IT problem that does not help the core business. Whether the technology was great or not, these vendors missed the fact that IT problems were solved by my lieutenants. At my level dealing with the rest of the C-suite and the board, my focus was ensuring that we could provide technology to solve the enterprise’s strategic initiatives or business problems and create business differentiation.

Technologies like AI and machine learning (ML) are not vaporware. When done right they hold the potential to unlock tens or hundreds of millions in new revenue or operational cost savings. But the key for its success is focusing on the strategic use cases that matter at the C-suite level, otherwise AI and ML can quickly join most initiatives that don’t even make it past the proof-of-concept phase.

Speaking from my work experience in CPG, the three most important initiatives determining the direction of the company are as follows:

  1. Shortening the time from production to reaching the consumers of the product
  2. Fulfilling consumer expectations for hyper personalization products
  3. Building a direct relationship with the consumers to get to know them.

I previously discussed the importance of agile machine learning in order to reduce the time from production of products to reaching the consumer of those products. In this article, I want to focus specifically on hyper personalization, which has immediate implications for how CPG brands manufacture products and their need to build a direct relationship with their consumers to know them well enough to personalize products specifically for them. 

Hyper personalization is a concept of knowing your customer well enough to create personalized products for them. This is not just personalizing the name of the product or branding the product with the consumer’s name or providing the delivery status of the product – but creating products based on the needs, likes and dislikes of the individual. For example, L’Oréal is experimenting with a service by which individuals can provide a picture (using their smartphone) of their skin. Based on the skin tone, types of skin oils, weather conditions in their geography and many such parameters, L’Oréal can produce a product specific to that specific consumer. This personalized product can then be shipped to the individual with status being provided at each step of the way.

In order to fulfill hyper personalization, CPG brands need three new capabilities in their organization. The first two are very logical. They are as follows

1)    The ability to cost-effectively produce individualized products and not just in masse

2)    The ability to take these products from the shop floor to the customer’s front step in an acceptable time frame (“acceptable” timeframe defined by the consumer).

Providing these two capabilities, requires the CPG manufacturer to predict the products that their customers will want so products can be created and start the shipping process almost before someone orders a personalized product. Most CPG brands have started collecting the types of data needed by selling to consumers directly. This way they have a sense for a customer’s preferences (colors, styles, etc.) and consumption patterns (how often they buy each thing). Along with other data (lookalike segments, weather, etc.) these brands can use machine learning to more accurately forecast demand down to the individual level at scale.

Typically, this capability requires many ML models to be created since there are many factors that cannot be digitized – like the impact of weather on the forecast of products, hence CPGs are ending up with thousands of models even for a single use case. Practically speaking, this requires MLOps that can deploy, manage, and monitor the ongoing performance of tens of thousands of models (here is a post that details how to scale your model operations from one model to potentially thousands).

For CPG brands to get AI and ML to deliver value that meets C-suite expectations, they need to think beyond building any single ML model. It means thinking about how to industrialize model operations to scale to the strategic challenges the company cares about. Hence, CPG brands need to consider bringing in MLOps technology in their environment. The combination of many ML models to create a use case and the implementation of MLOps creates business value at the CxO level and hence may cross the proof-of-concept stage.

About Manish Sinha

Manish is a special advisor to Wallaroo and an award winning CIO with global and multi-industry experience. Winner of the CIO of Year and the Peer CIO awards. Nominated by peer CIOs to the Wall Street Journal CIO Forum in USA. Member of the Senior Leadership Team in IT for Loreal, ANSYS, UBS, Yahoo and Microsoft. Broad range of experience from Infrastructure, Artificial Intelligence (AI), Cloud solutions to Business applications. Recognized by PwC and ANSYS Board for making multi years of information security enhancements in 18 months. Used AI, Master Data Management (MDM) & Enterprise Monitoring to predict incidences and prevent outages. Active in giving back to the IT community through advisory councils (ServiceNow & Microsoft), VC advisory forums, writing publications and speaking at global conferences like Economic Times, and National CIO Review. 

Table of Contents



Related Blog Posts

Get Your AI Models Into Production, Fast.

Unblock your AI team with the easiest, fastest, and most flexible way to deploy AI without complexity or compromise. 

Keep up to date with the latest ML production news sign up for the Wallaroo.AI newsletter

Platform Learn how our unified platform enables ML deployment, serving, observability and optimization
Technology Get a deeper dive into the unique technology behind our ML production platform
Solutions See how our unified ML platform supports any model for any use case
Computer Vision (AI) Run even complex models in constrained environments, with hundreds or thousands of endpoints