The answers from Stéphane Roder, author of “Practical guide to artificial intelligence in business” (ed. Eyrolles).
Stéphane Roder has headed the AI Builders consulting firm since 2018. Here he describes the effects of this new technology for the world of franchising
Should the world of franchising also be interested in artificial intelligence?
The franchise universe has two characteristics. The first requires the franchisor to market the brand, and therefore to optimize its costs. The second involves being the supplier of the product to its franchisees. There will then be a problem of flow, to be maximized, whether it concerns purchases or stocks.
In terms of marketing, tools based on artificial intelligence can firstly improve the return on each euro invested by 20%. Then, when it comes to purchasing and the supply chain, these tools save time and get to the point.
Why are we talking so much about artificial intelligence today?
Because we are lucky to have a publisher, Microsoft, which has remarkably “marketed” the release of one of its technological assets, OpenAI! Many people have used it, and they really liked it. The tool impressed. Because ChatGPT has almost human behavior, close to the Turing test, that is to say the machine has the ability to imitate human conversation. But this development is not a revolution: in fact, we now have more efficient architectures which allow the emergence of these generative AIs.
Why should a franchise professional define an AI strategy?
Those who manage a business know: margin matters. We must therefore seek to improve our costs. In this way, a company is sustainable and can develop. There is therefore a real challenge for the franchisor: he must have perfect optimization of his investments and his income. For franchisees, it will depend on their business, some will be able to benefit from AI, others not. But the former will need to provide the latter with the tools to optimize their work. Otherwise, the main risk is having a heterogeneous network that does not function well. The franchisor therefore has an interest in providing the right use cases on the shelf or integrated into the workstation. It is up to him to question the solutions he can provide and to target the points where AI will have a positive impact.
Are the tools based on this technology mature enough to be taken into account in a strategy?
We see, for example, that some, based on generative AI, encounter hallucination problems.
Let’s not forget that ChatGPT and generative AI represent only a very small part of AI. Companies have been using machine learning since 2012 in many areas, notably to make qualitative and quantitative predictions, which they use in inventory management. But you are right about the reality of these hallucination problems at the level of generative AI, which we are beginning to understand and work around. However, not all of these are intended to be integrated into company plans. It is above all a fashion, and the general public must always bear in mind that AI corresponds above all to applied mathematics, whatever the prediction or the tool.
Should a franchise manager evolve his organization to better understand AI?
From the moment a franchisor offers this technology, its response must be industrial, which involves setting up an ad hoc organization to define use cases, in particular by calling on consultants or skills. internal. Then, you have to do development and create a data office. At this stage, it is not the size of the franchise that matters, but the number of use cases.
What are the pitfalls to avoid?
You should not implement AI that does not serve the company’s objectives. In other words, do not launch into this field to say that your company is at the cutting edge, nor absolutely want to put it everywhere. Today, everyone wants generative AI. But it should only be used if it contributes to the development of your society and is not a useless investment.
You say we need to demystify AI, why?
AI can be very useful. It’s a powerful tool that brings a lot of value, but it can’t do everything, and has its own limitations. Despite the extraordinary marketing strength of Gafam, we must not let ourselves be carried away by the wave around generative AI. For example, we tell our customers: “In your company, don’t make Copilot available to everyone, first look at who it will really benefit.”
After the major language models, what should we expect? What trends do you anticipate?
We are extraordinarily lucky. ChatGPT has enabled a huge boom in the sector, with money invested in companies and research centers. There is a real innovation-industrialization-generalization cycle that is starting before our eyes.
Beyond this first observation, we observe a transition from large language models (LLM) to smaller ones. Using 175 billion parameters like GPT-3 in each query is too expensive. Hence the arrival of new, more modest software with 2 billion parameters, very specialized and vertical. These small specific models will be able to equip workstations. If franchisors want to invest in AI, they will have no choice. They will have to do it in an industrial and well-designed way. Conversely, not being interested in it means taking the risk of no longer being competitive and leaving the market. And be careful! these radical changes can be violent.
Our summary in 5 key points by L’Express Connect IA
(verified by our editorial staff)
Here is a five key point summary of the article on franchising in the age of artificial intelligence.
Importance of AI for Franchises: According to Stéphane Roder, franchises must adopt AI to optimize marketing, purchasing and inventory flows. AI can improve the return on marketing investments by 20% and simplify operational processes, providing a crucial competitive advantage.
Strategy and integration: Franchisors must define an AI strategy to ensure investment and revenue optimization. Providing AI tools to franchisees is essential to ensure network homogeneity and avoid performance disparities between different points of sale.
Maturity of AI tools: Although some generative AI like ChatGPT may experience hallucination issues, AI technologies have been widely used since 2012 for qualitative and quantitative predictions. It is essential to understand and work around current limitations for effective use.
Avoidance of pitfalls: It is crucial to only adopt AI if it serves business goals, rather than following a trend. Investing in AI without a clear business benefit can lead to wasted investments.
Future AI Trends: The future of AI involves smaller, specialized models that are economically viable and can be integrated into workstations. Franchisors must invest in a structured and industrial manner to remain competitive, otherwise they risk losing their position in the market.
