Generative AI is seductive. The remarkable pace of adoption and the natural language simplicity of its interface has quite rightly grabbed attention. The potential business use cases are wide-ranging and, in many cases, profound but we are still very early in the adoption cycle. Firms looking to use Gen AI or other disruptive technologies to improve products or services need to consider their options before rushing in. Broadly speaking there are three main avenues available:
- The first is look for existing IT suppliers to innovate. Many SaaS platforms are experimenting and releasing new AI driven features. These new concepts can be tested with discrete groups of customers and distributed to 100% of users once refined. Multiple marketing SaaS platforms offer content generation based on individual customer preferences rather than segmented models using ‘Content Assistants’. This allows business to ‘use Gen AI’ within the context of a platform they already understand. There is little to no need to develop any code or even understand how it works. There are limitations in scope but as a first foray into an uncertain world this approach offers an opportunity to test and learn in a familiar environment with strong in-built guardrails.
- A more adventurous second approach is to use dedicated AI powered tools to address specific business issues. Stepping away from mainstream SaaS platforms and exploring opportunities to deploy more bespoke technologies, often in combination. A more business-focused approach such as this means that firms will be focusing on issues that matter and then looking for tech to help them resolve a problem. The breadth of potential use cases is much wider but at the same time the implementation complexity and cost will increase. For example, new product use cases can be better informed and connected with Gen AI. Specific features and functionality can be more customer-centric by analyzing customer feedback, social media data, and product reviews. Once established these capabilities can be deployed routinely and form a regular part of the product development lifecycle, providing early adopters with distinct competitive advantages in getting to market faster, and reducing the cost of customer listening. Similar opportunities exist within a service context with conversational chatbots already being used across many industries. Predictive analytics capabilities are far more sophisticated with GPT using elements of machine learning to predict customer reactions in ways that were only previously available to large corporates with dedicated AI tech teams. With Large Language Model (LLM) applications being released on a regular basis there is arguably too much choice. Firms using the business issue first approach are capable of solving more complex problems but will also need to proceed with more caution. Sharing potentially sensitive customer data with 3rd party applications represents an obvious risk that will need to be managed carefully. Integrating an AI platform with existing business systems and processes requires careful planning, testing, and monitoring.
- The third approach is more bespoke, data-first, and offers the greatest ROI. Every firm is different and will generate different datasets that uniquely describe how a business is run. Capturing and organizing data across the business is the foundational component of any AI/Disruptive tech capability. Much of the data may be unstructured such as voice or images but all of it is potentially valuable. The breakthrough technology in this case is the AI Data Pipeline. All three major cloud platforms (Azure, GCP, and AWS) offer sophisticated data management and AI solutions that are all cloud-based. They offer very high-grade security with an increasing range of sophisticated AI and data management tools. These AI tech building blocks open the door to more bespoke and potentially differentiating capabilities. At one end of the scale firms can develop their own AI algorithms and deploy them directly to work online or internally. At the other end, their options for leveraging a commercial AI solution are increased with the ability to connect with their own proprietary data. In all cases firms can be confident that they are improving their corporate knowledge and AI skills. Data pipelines and AI platforms will continue to evolve and become commoditized eventually. The firms that will succeed in the medium term will be those that are best able to execute in bringing AI and disruptive tech to their business.