Written by Mathieu Louis
With the launch of ChatGPT at the end of 2023, companies' awareness of the potential of AI has increased significantly. Over the past year, we have observed companies struggling to get these new opportunities right. A common impulse is the desire to solve all problems with generative AI capabilities. While exciting, is generative AI the magic wand that will solve all your problems?
Challenges to address?
Generative AI can indeed be a tremendous source of innovation and added value for your teams when used appropriately. However, before jumping into implementing generative AI-powered solutions, it's important to understand the challenges you're trying to address. As we help clients examine the root of their problems, we often find that simpler solutions can deliver better results at a lower cost.
A recurring example is the desire to use generative AI for classification tasks. Not only are the results less likely to be convincing, but managing the generative AI classifier in production is more complex and expensive than traditional machine learning methods. Companies thatdon't have their own data to use can quickly benefit from generative AI-based classification methods, but it's essential that they invest in structuring their data and making it ready for traditional machine learning and analytics to maintain their competitive edge. One solution we suggest, if companies don't have enough structured data, is to help them generate synthetic data when it's suitable for the task.
Common use cases
Two common use cases where we see generative AI is helping teams:
Speed up tasks involving understanding and creating language: Whether it's summarizing content, writing standard code, reporting from textual data sources, or manipulating text to produce specific outputs, generative AI can streamline these processes. One of our most popular examples is a code generation assistant we created. While freeing up developers' time to think more about the structure of the application and the algorithms to be written, the code generator wizard provides all the coding blocks and also creates test scenarios.
Enhance creativity: As a partner in mock-up creation, brainstorming and strategic planning, generative AI can act as a catalyst for creative thinking, challenging ideas and proposing new plans. One concrete example is a mock-up builder we've deployed for a client that increases productivity during brainstorming sessions by creating quick and realistic mock-ups while also suggesting new ideas.
Define scope & capabilities
When implementing generative AI solutions, it is crucial to understand the required data inputs and expected outputs. How can quality be controlled? Who will use the system, and how will they use it? Taking the time to clearly define the scope and capabilities of the system, while ensuring that employees are equipped to use these new capabilities, is critical to the success of such projects. We have found that even if we provide GenAI assistants with incredible capabilities, most employees will not use them if they are not supported in this transition. Customised Generative AI tools will always need guidance when set up, so the impact of data and human input alongside change management cannot be underestimated.
As we did when we introduced our productivity booster Remus chatbot to our teams, we always help our customers in creating an implementation plan that takes these aspects into account.
BrightWolves
At BrightWolves, we work with our clients throughout the implementation process of a generative AI project. From the initial stages of defining the scope, teams and technologies to use, through to implementation and supporting the teams in using these new capabilities. Our expertise in AI can help accelerate your digital & data transformation by providing valuable guidance on best practices and implementation strategies.
If you want to know more, do not hesitate to reach out to our AI expertise team Olivier De Moor, Bjarne Keytsman, and Mathieu Louis.
Comments