From Cost-Saving to Cost-Sinking: The Pitfalls of Unstrategic AI Adoption in Retail
We are in an era where the allure of generative artificial intelligence (GenAI) in retail paints a futuristic picture of efficiency and innovation, the rush to adopt this promising technology can sometimes lead to patchwork processes and customer experiences that end up requiring rework and/or optimization. The promise of GenAI to streamline operations, personalize customer experiences, and ultimately reduce costs is undeniable. In the retail sector alone, GenAI could deliver value equal to an extra 230 to 390 billion dollars per year in terms of productivity if all use cases were implemented. Yet, as many forward-thinking leaders in the retail sector are discovering, the journey towards AI-driven transformation is fraught with complexities far beyond the initial allure of cutting-edge tech. Without a coherent digitalization strategy, the vision of AI as a cost-saving savior can quickly morph into a cost-sinking quagmire, draining resources, time, and potential with little to show for the investment. In this article, I will try to share plausible strategies and adoption scenarios for GenAI in retail and highlight the cost-sinking pitfalls.
Laying the Foundation for Successful GenAI Adoption
Venturing into the GenAI adoption for a retail company is akin to mastering the art of chocolate making. Before a chocolatier pours the first batch of fine cocoa into molds, there’s a period of preparation that ensures the chocolate will be sublime. In the same vein, C-level executives must first prepare their enterprise by fostering a workplace that understands and embraces the change GenAI brings, much like a chocolatier ensures the quality of beans and the readiness of equipment. First and foremost, it is imperative for an organization to align their workforce to co-exist with potential GenAI applications. Priorities within different teams and third-party vendors could create resistance; imagine a retail company that is working with a well-known marketing agency and a success story driven marketer that is working in the marketing department. A typical GenAI use case in marketing is content generation for campaign banners. Utilizing this use case within this scope means a reduction on the budget of the marketing agency and less playground for the marketer. The agency will try to resist their budget reduction while the marketer working for the retailer feels threatened about losing his/her job to a middleware solution. Secondly, deciding whether to rent, buy, or build GenAI solutions is not unlike a chocolatier’s decision on sourcing their chocolate. Renting technology is like ordering pre-made chocolate from a renowned supplier — quick and easy, but with less control over the flavor profile. Buying is like purchasing cocoa beans from a trusted farm to process in-house, offering more control and a signature taste. Building, however, is the equivalent of growing your own cocoa trees from scratch — a long-term commitment that promises a unique product truly reflective of your brand. Then, there’s the matter of the company’s digital maturity, which can be likened to the chocolatier’s expertise. It’s not enough to have the finest Ghanaian beans if you lack the skills to temper chocolate correctly. A retailer must assess and cultivate their digital capabilities to ensure they can support and enhance GenAI’s sophisticated algorithms.
There is a good visual representation of GenAI prioritization in a retailer (1). Top division specific use cases need to be supported with continuous foundation layers that are discussed above. So, in this sweet narrative, the essence remains clear: embracing GenAI in retail is not just about adopting new technology; it’s about the meticulous blend of preparation, strategic decision-making, and the nurturing of digital expertise — all essential ingredients in the recipe for a successful GenAI transformation, as indulgent and sophisticated as the finest chocolate.
Upgrading Your Human OS: Embracing GenAI in Retail
To navigate the GenAI landscape, workforce readiness needs to reach a certain threshold and develop continuously with the development rate of AI technology. GenAI applications will get increasingly more expensive as top enterprise solutions will reach to a certain customer base and start targeting for more profits. This was the case with the streaming platforms. They started with relatively cheap monthly fees to attain more customers within their early adoption phase. When small time platforms merged with big companies like Netflix and Disney and platforms gathered certain customer bases, they started to ramp up their subscription prices to fund their content generation and attract more investors. Same pricing curve could be expected from SaaS based GenAI services. Investing to these GenAI solutions without the following steps could drastically decrease the efficiency gains.
1. Assess workforce impact: Utilize a skill gap analysis to identify current skills of your workforce and compare them with the skills required for future roles created or transformed by GenAI. Transform existing roles and/or create new roles to align your business needs for GenAI adoption. Some roles’ monotonous tasks will completely be replaced by AI solutions while other roles will be augmented.
2. Develop a training program: Design GenAI tailored upskilling and reskilling trainings to help employees develop themselves for new or evolved roles. Foster a continuous learning culture that encourage personnel to adopt new GenAI solutions into their business operations.
3. Redefine KPI’s: As roles evolve with GenAI, update key metrics to reflect new objectives, ensuring these metrics are aligned with the GenAI expectancies from shareholders will create a more permanent adoption.
Rent, buy or build: Defining adoption approaches
In the pursuit of incorporating GenAI into our retail strategy, we face a strategic decision: should we rent, buy, or build our solution? Renting offers immediate access with less capital outlay, akin to a short-term lease that’s ideal for testing the market’s waters. Buying a ready-made solution is an investment, like acquiring a turnkey store; it provides a quicker start than building from scratch and often includes vendor support. Building our own solution, however, is the bespoke suit of options, tailored to fit our unique business needs, but it requires significant time and resources. Each path has implications for control, cost, and speed to market. Our decision will hinge on a clear-eyed assessment of our needs, capabilities, and long-term vision, ensuring we choose the path that aligns with our strategic goals and delivers the best return on investment.
Risks can be mitigated by analyzing each use case by their own merit and prepare a ROI analysis. Different initiatives could be taken for different strategic needs (2). If the need is to assist software developers with more day-to-day coding practices, a Taker approach can be chosen, if we are looking for a customer service chatbot fine-tuned with sector specific knowledge and chat history, a Shaper approach is more appropriate. If the need is to create a foundation model for assisting in patient diagnosis, a Maker path can be taken. The bottom-line is that there is not a single turnkey solution for developing GenAI use cases. In some cases, it is simply smarter to rent a tailor-made problem solver GenAI software to see immediate effect while it is more beneficial to create a foundation model for a company specific need to create a competitive advantage with GenAI application.
Utilize the Right Solution Stack for Maximum Effect
When the digital maturity of the company is high enough for GenAI solutions to be adopted and the necessary approaches are studied for each business case, the final step could be to find the right application areas for each business operation to maximize the value output in a feasible time frame. As the GenAI solutions developed rapidly in the market, a medium sized project that can be deployed in 6 months can be considered a lot of time. Noting the finite resources of a standart retailer company, green lighting and placing wrong projects first in line could result in a failed business year for the utilization of GenAI.
To extract the most value within a year, each GenAI use case can be analyzed according to company and local market data to find the optimum use cases. Gartner’s use case prism for generative ai in retail (3) is a good example for quantifying each application areas in terms of their value and feasibility. With the limited resources of a retail company, committing to a valuable project such as customer behavior modeling is a tough start. These types of big investments should be considered when company’s AI adoption is high, and all the quick wins are already utilized.
Projects within company’s GenAI portfolio can be methodically evaluated and classified based on two key dimensions: value and feasibility. The value dimension encompasses the enhancement of operational efficiency, risk management capabilities, and non-financial benefits such as customer satisfaction and brand reputation. On the feasibility side, there are technical viability, the preparedness of the internal infrastructure, and the alignment with external market conditions.
Company’s focus should be directed towards those use cases that rank high in both value and feasibility — the “Likely Wins.” These use cases promise the greatest immediate impact and the smoothest implementation, representing an optimal allocation of our finite resources. This strategic approach, as depicted in section (4), ensures that companies can prioritize initiatives that are poised to deliver significant returns, both in the short and long term, solidifying the market position and fulfilling commitment to innovation and stakeholder value.
In wrapping up, the success of GenAI in retail hinges on strategic planning and digital know-how. Retailers who plan carefully and understand GenAI’s complexities will be the ones to benefit most. The goal is to adopt GenAI smartly, ensuring it brings real value and helps the business grow. Let’s focus on these smart choices to make the most of what GenAI offers.
Resources
https://www.gartner.com/en/doc/use-case-prism-generative-ai-for-retail
https://blog-idceurope.com/how-retailers-and-brands-are-taking-advantage-of-generative-ai/
