Printable version here

Introduction: companies are far from done with digital transformation and now comes AI

For over a year, gen AI has been on everyone’s lips. Boards are pressing their CEOs to have a strategy for incorporating AI in the business even as many executives still don’t know where to start. The IADS organised a conference with Bain & OpenAI as early as July 2023 to explore the topic. In any case, 90% of commercial leaders expect to utilize gen AI solutions “often” over the next 2 years. They are cautious though, and are most enthusiastic about use cases in the early stages of the customer journey including lead identification, marketing optimization, and personalized outreach.

This article delves into AI developments that have the potential to impact and improve retailers’ operations, in order to define a non-exhaustive list of existing use cases already implemented. In CRM and marketing, gen AI helps to better target audiences and shift towards ultra-personalization. The impact of conversation tools has already been visible in copywriting and chatbot developments as AI has been influential in boosting creativity. When it comes to sales functions, AI tools have the power to increase sales thanks to better and tailored customer experiences. In terms of supply chains, AI has not been fully developed, but there is a lot of potential. Finally, AI has already impacted HR practices.

Gen AI for CRM and marketing: advancing towards ultra personalization

Enhanced audience targeting

Gen AI can combine and analyse large amounts of data (demographics, customer data and market trends) to identify additional audience segments which may have been overlooked in existing customer data. Gen AI can significantly reduce the time spent researching and creating these unique audience segments. Without knowing every detail about these segments, gen AI tools can automatically propose tailored content such as social media posts. Then, marketing (collaborating with sales) can use gen AI to create sales campaigns to reach prospects. This requires efficient data management: a comprehensive and aggregated dataset is needed (such as an operational data lake pulling in various sources) to train a gen AI model that can generate new audience segments and content.

Transitioning from personalisation to hyper-personalisation

Initially, personalization was limited to traditional market segmentation like gender, age and income. With AI, personalization has become more sophisticated, allowing retailers to understand and anticipate customer preferences more precisely and create tailored experiences able to foster loyalty. In that regard, AI technologies including deep learning and machine learning (ML) are used to analyze structured and unstructured data to create a complete view of each customer. Alix Partners sees the most significant AI potential in combining gen AI with ML to identify high-potential customers (based on customer lifetime value) and determine which ones are likely to make additional purchases. Together, ML can analyse complex data to identify patterns while gen AI can generate new content. This approach paves the way for real-time, highly personalized omnichannel experiences. For example, companies like Stitch Fix (online personal stylist) use gen AI to interpret customer feedback for product recommendations. The next step for companies is to transition from reactive to proactive personalization, providing 100% individualized content across channels. Here, the challenge will be about ensuring ethical data usage while protecting customer consent, privacy and security. Customers are increasingly expecting personalised experiences and relationships with their favourite retailers: as discussed during the IADS Operations Meeting dedicated to Chief Customer Officers in November 2023, the best way is to be as transparent as possible with customers and explain why and how their data will be used to help them get what they want.

AI conversation tools provide solutions for copywriting and chatbot

Gen AI’s capacity for producing natural-sounding language makes writing one of the tasks it’s wellsuited for. In addition to ChatGPT, start-ups like Jasper and Hypotenuse offer new tools. Jasper scales up marketing content like blog articles, social media posts, sales emails and website copy. By providing a few keywords, Hypotenuse users will instantly turn them into full-length articles and marketing content. On their side, tech providers such as Shopify, Salesforce and Amazon are adding gen AI copy options to their platforms to help companies streamline the writing of everything from marketing emails to product descriptions. In that regard, during the 2023 IADS Operations Meeting dedicated to Chief Customer Officers, El Corte Inglés noticed that AI product description is sometimes better than when done by people. While human oversight is needed, AI copywriting tools can already automate laborious and mundane work.

AI-powered chatbots (especially useful for platforms with a vast inventory) can enhance the shopping experience by understanding and responding to natural language and offering tailored product suggestions. Importantly, they create an iterative experience in which shoppers can respond to the results with feedback or additional questions, guiding the bot towards what they want. However, bots face limitations in providing accurate product suggestions and require a deep understanding of the retailer's inventory. Also, BoF made tests in spring 2023 and found that bots' replies can sound automated. So far, the best solution is to complement traditional search with a gen AI assistance.

A larger goal for many fashion players is to use customer data to personalise chatbot’s responses. The bot could use the data to offer specific sizes based on a customer’s preferences, for example. It’s an ambitious goal, though, and requires brands and retailers to have their customer data at hand and to be able to map it to their inventory. Google research showed that 46% of organisations think gen AI can address shopper enquiries with interactive responses beyond just product recommendations. Also, 43% want to use it to analyse emotional sentiments in customer feedback. When it comes to IADS members, Galeries Lafayette and Manor are currently fine-tuning their chatbots.

AI can boost creativity

It’s not a magic wand, but AI can support fashion design

Tools like DALL-E 2 and Midjourney have made it easier to create fashion content through generative AI. Whether it’s for branding purposes or to truly create design variations, some brands are already leveraging generative AI for product design. It provides designers an easy way to design countless variations of a piece of cloth, mixing inspirations to see what the outcome might look like. It does come with challenges. AI-generated designs still need manual edits and integrating the process into existing workflows can be difficult because it doesn’t consider real-world factors like fabrics and construction. Designs still generally require manual editing with separate software (for example to change a colour). Despite companies working on 3D gen AI, images are two-dimensional for now: the design only shows the front of an item, leaving the designer to create the rest of the garment. Finally, AI can produce concepts that are difficult or impossible to construct, making it impossible to translate them into finished products. Finally, intellectual property issues exist with AI-generated designs. Nonetheless, gen AI can represent a powerful tool to boost creativity which could help Private Labels design teams for example.

Generating unprecedented visual content

Visual content has emerged as another promising use of gen AI for fashion brands and retailers, which are under constant pressure to renew visuals for marketing, social media and e-commerce. Gen AI has the potential to provide more creative freedom and shorten production timelines as scouting locations, finding models and styling them are no longer necessary. For instance, Casablanca fashion brand used AI to produce stylized ad images, demonstrating AI's potential to revolutionize content creation. As AI-generated images are rapidly developing, Galeries Lafayette created amazing AI interpretations of its famous cupola. Besides unlocking additional creativity, using AI can offer cost savings and creative flexibility but may also impact traditional roles in image production. Also, there are potential sustainability benefits since AI eliminates the need to travel to shooting locations and reduces waste (multiple samples and sets discarded after use). Recent Google research shows that 39% of the surveyed organisations use gen AI to empower creative retail teams to curate bespoke images and creative content for campaigns and editorial placements.

To what extent can AI help develop sales?

AI to enhance customer experience and sales…

With its ability to analyse customer behaviour and preferences, gen AI can assist with hyper-personalized follow-up emails at scale. When thinking about clienteling, it can also act as a virtual assistant for each sales associate, offering tailored recommendations, a warm welcome to new customers, and reminders and feedback, which can each result in higher conversion rates. As a potential sale progresses with a customer, gen AI can provide real-time guidance and predictive insights based on an analysis of historical transaction data. Finally, AI can boost sales performance by automating mundane sales activities, allowing sales associates to spend more time with customers and leads (while reducing the cost to serve). The potential applications of gen AI and ML extend further, including matching customers with relevant sales associates. The integration of these technologies can significantly enhance outcomes, making it a promising investment for retail businesses. Some companies that are empowering this process are BSPK, Clientela, and FindMine.

As announced during CES and NRF in January 2024, Walmart's strategy is going big on AI with many different use cases proposed, showing the width of potential applications. One of the initiatives aims at making sure people’s refrigerators are always stocked. Using the example of a party a customer would throw for the Super Bowl, Walmart explained their AI-powered app will show everything people might need instead of having them search for chips, drinks or a new large-screen TV. Also, as explained during the third IADS CEO quarterly exchange of 2023,https://www.iads.org/web/iads/5747-iads-ceo-meeting-3.php Cyrille Vincey (Partner, Advanced Analytics and Retail practice, Bain & Co) explained how AI helps Carrefour in developing sales. For example, online shoppers can ask the chatbot for ideas for meals for a family of 4 for a week. In response, the chatbot provides recipes and translates them into a bucket list, and ultimately into a full basket. On its side, El Corte Inglés just launched an online ChatGPT personal shopper giving fashion advice and able to increase the conversion rate and hopefully the basket size. Overall, it has been an excellent learning experience, and the first results are promising.

… And reduce returns by solving fit issues 

Amazon Fashion has introduced new AI-driven features to address the fit problem in fashion e-commerce. The new tool aims to reduce returns and improve the overall shopping experience. The personalized size recommendations algorithm evaluates sizing relationships between brands, reviews, and customer fit preferences to recommend the best-fitting size. The AI-generated fit review summarizes customer feedback, helping shoppers make informed decisions about sizing. Also, Amazon has improved its size charts using AI to enhance accuracy and consistency, making them easier to follow and potentially addressing the variability in sizing systems across styles and brands.

AI-driven startup solutions addressing fit issues are developing quickly. 3DLook, a guest speaker during the 2023 IADS Operations Meeting dedicated to CIOs and CTOs explained how they help brands such as Bershka increase revenue and cut costs related to fit and sizing problems. AI and 3D engines deliver advanced body measuring technology for more intelligent fit experiences.

From forecasting demand to sustainability, AI has not fully transformed supply chains yet

Why do companies struggle with using AI for supply chains?

Companies have struggled to use AI to address fundamental supply chain challenges. Supply chain management is complex as it requires the participation of several functions (including procurement, manufacturing, logistics and sales) and sub-functions (such as demand planning, inventory planning and scheduling). Besides, organizational structures and incentive systems motivate employees to optimize the performance of their own function or subfunction rather than the end-to-end supply chain.

Companies often try to improve their supply chain performance by adding more people to a function. But the problem is typically a lack of knowledge, which cannot be solved simply by creating larger teams. High-potential performers often do not regard supply chain management as a preferred long-term career path and move to other functions after only 2 or 3 years. Because of the high turnover, institutional knowledge ends up dispersed across the company or escapes the company altogether.

The root cause lies not with technology but with how and where companies are applying it. Probably because they consider it is too risky, companies have not pursued the more valuable application of using AI to make recurring decisions by recognizing patterns in big data that humans cannot see.

It is too soon to rely on AI to predict what shoppers will buy

There are challenges in using AI to predict what shoppers will buy. Experts are cautious, emphasizing that the emotional nature of fashion purchases represents a significant hurdle. AI should be seen more as a support tool for experienced merchandisers rather than a replacement for human expertise. AI-driven services, such as in-season reorders and pricing optimization, are gaining traction though. Yet, brands are sceptic about AI-powered demand forecasting even though leveraging ML can provide more accurate predictions and allow inventory optimization by analyzing historic demand, supply data and trends. Although AI has the potential to provide more precise forecasts than historical methods by considering a multitude of variables, concerns remain about its readiness to entirely replace human decision-making.

Supply chain automation

Modern supply chain automation is not possible without AI. AI gives supply chain automation technologies such as digital workers, warehouse robots, autonomous vehicles, etc, the ability to perform repetitive, error-prone tasks automatically. Thanks to AI, automation can be fulfilled in the back office (document processing), logistics (companies like Amazon are investing in autonomous trucks), warehouse management (Ocado), quality checks and inventory management (thanks to AI-enabled computer vision systems).

Improving sustainability

Sustainability is a growing concern for supply chain managers since most of an organization’s indirect emissions are produced through its supply chain. AI can help improve supply chain operations to make them more sustainable. AI-powered tools can help optimize transportation routes by considering factors such as traffic, road closures, and weather to reduce the number of miles travelled. For instance, DHL uses AI to optimize vehicle routes and reduce fuel consumption, resulting in lower emissions and improved sustainability. Since AI-powered forecasts should help maintain optimal inventory levels, carbon emissions attached to storage and movement of excess inventory could be reduced.

Supporting functions: financial forecasting and transforming HR

AI for financial forecasting

After marketing, financial forecasting is the second area where retail and CPG executives will invest in AI tools. Algorithms can analyse large amounts of financial data and generate forecasts based on historical trends, market fluctuations and other factors. AI financial forecasting can also be used for a variety of purposes, such as predicting stock prices, forecasting economic growth, identifying potential investment opportunities and make better decisions about inventory management or pricing strategies.

From recruitment to job performance and professional growth

There is a growing use of AI in crafting job postings, shortlisting candidates, matching applicants to job ads and personalising communication with applicants, but also in specialized tasks like predicting the candidates' future performance. Companies like Skims are using AI platforms like Dweet for recruiting, which shows promising results in broadening candidate searches. On its side, Eightfold AI's platform is designed to predict the future roles an employee might be good for.

In that regard, Gen AI can help identify career paths and opportunities for employees, hence facilitating a more personalized career development journey. It can be particularly beneficial in visualizing career trajectories and identifying potential role models within an organization. In the future, AI could assist in identifying candidates for promotions and better role placements, reducing talent attrition costs. Also, the technology can be used as a productivity aid in performance reviews. It can assist in creating initial drafts of reviews by synthesizing feedback from multiple sources, thus allowing managers to focus more on individual development and growth.

Overall, gen AI could streamline administrative tasks for HR to focus more on strategic aspects of talent management. The technology should be first used to improve decision-making and performance management.

Navigating bias in recruitment

However, the success of AI in recruitment still heavily relies on human oversight to address its limitations and potential biases. The technology aims to reduce biases in hiring, but its early development stage means it's not fully reliable yet. For now, experts say AI could even amplify existing biases related to age, gender, and race. AI is not advanced enough yet to completely replace human involvement in hiring. However, some tools aim to create a "positive bias" by focusing on desired skills rather than disqualifying candidates. Dweet's software, for instance, highlights candidates with relevant experience and doesn't penalize gaps in resumes or lower educational levels.

Conclusion: first comes a clear vision for AI

Looking ahead, the potential of Gen AI in retail and fashion is immense and still unfolding. But for now, top executives are admitting they're far from ready to deal with changes brought by generative AI, according to a new global survey by Deloitte's AI institute. The problems may only get worse. Executives who reported the most investment and knowledge in generative AI capabilities are the ones most worried about the technology's impact on their businesses. Only 1 in 5 executives believes their organization is "highly" or "very highly" prepared to address AI skills needs in their company. Only 47% say they are sufficiently educating employees about AI. The majority of executives said their organizations were focused on the tactical benefits of AI, such as improving efficiency and cost reduction, rather than using it to create new types of growth.

At successful companies, McKinsey found there is a clearly defined AI vision and strategy. Also, more than 20% of digital budgets are invested in AI technologies. Teams of data scientists are employed to run algorithms to inform rapid pricing strategy and optimize marketing and sales. Finally, strategists are looking to the future and outlining simple gen AI use cases. Such trailblazers are already realizing the potential of gen AI to elevate their operations. Players that invest in AI are seeing a revenue uplift of 3 to 15% and a sales ROI uplift of 10 to 20%.

While the application of gen AI in a retail business can seem overwhelming as it can fit into almost any piece of the business, the earlier that it is ‘plugged in’ is better as it only gets wiser with time and data. The time to jump in the AI train is now.

Credits: IADS (Christine Montard)