Delivering Gen AI value through IBM Client Engineering
In a time of constant innovation, generative artificial intelligence (Gen AI) is quickly becoming one of the most defining technologies of our time. This may sound extravagant, but there’s no denying its great potential to increase productivity and efficiency.
Gen AI's potential spans diverse applications, including enhancing customer service in telecom by leveraging natural language processing for quicker issue resolution, streamlining retail and supply chain through product finders and cost/revenue mapping, automating document redaction in legal/public sectors, and optimizing manufacturing with visual inspection and part identification.
Despite its potential, businesses struggle to identify potential use cases for Gen AI for their respective business and industry. Off-the-shelf Gen AI tools do not always bring the value customers seek when dealing with specific challenges. To ensure business-specific or industry-specific challenges are tackled and real value is delivered through technologies such as Gen AI, IBM launched IBM Client Engineering, focusing on co-innovation and co-creation by working closely with customers and partners.
IBM Client Engineering
IBM Client Engineering’s teams consist of solution architects, developers, designers, engineers, and business technology leaders who, through different stages, including a workshop, technology discovery session, co-creation, and a playback session(feedback), bring to life a minimal viable product in just 4-6 weeks. The type of engagement focuses on anything from Gen AI strategy, skills development, or a solution for specific challenges while others want IBM to assess their business and showcase how Gen AI can improve their processes and operations.
Over the last six months, based on over 60 customer and partner use cases, IBM has claimed that the public sector, telco, automotive, CPG, and retail offer great potential and interest for Gen AI applications. IBM’s AI and data platform, Watsonx, is one such technology, but IBM Client Engineering is not restricted to IBM technology. Depending on customer needs and the company's existing IT estate, technologies from Microsoft, AWS, Google, Salesforce, Adobe, and many others are used to provide the best possible value-added outcomes for clients.
In a perfect world, bias would be removed from the equation, and the best suitable technologies would be deployed to deliver the highest value to the customer. However, it is often the case that proprietary technology is chosen, particularly if the customer does not show a strong preference to adopt a partner’s technology. Simultaneously, the adoption of different Gen AI tools from different providers is just as valid as multicloud adoption.
For example, in September 2023, UK bank NatWest expanded its collaboration with AWS to accelerate the use of Gen AI. In a co-creation effort, NatWest data scientists and engineers worked closely with specialist teams within the AWS Generative AI Innovation Center to develop responsible AI products on top of foundation models (FMs) available through Amazon Bedrock. Use cases include fraud detection through unusual payment patterns or tailored banking experiences such as personalized communications, financial education, products, and services.
Not long after, in December 2023, NatWest also started a four-week collaboration and co-creation on Gen AI with IBM Client Engineering, leveraging IBM Watsonx. This first proof of concept (PoC) focused on re-designing Cora’s interaction model (NatWest’s digital assistant), creating a ‘Natural Conversation Framework’ to strengthen Cora’s natural conversation style. The second PoC focused on transforming Cora into a trusted, safe, intuitive ‘financial coach’, leveraging Gen AI innovations. According to the NatWest lead for innovation of Cora, engaging with IBM Client Engineering in the same room not only enabled both teams to create solutions at speed but also changed NatWest’s approach to innovation. A culture of experimentation, collaboration, agility, and speed are now embedded into its innovation practice.
Risks and measurability
As Gen AI demand and initiatives continue to grow exponentially, it is not just about creating solutions for certain challenges. The risk and explainability of developed solutions must be considered. For example, Gen AI tools developed to provide quick and high-quality responses to customers, constituents, etc, must have safeguards in place such as explainability, i.e., why is it giving a specific answer, and is it factually correct? Another point to consider is measurability, i.e., does the tool deliver the expected value and outcomes set out at the onset?
Other risks include rubber stamping, i.e., as Gen AI adoption becomes more pronounced, will it make employees lazy because they blindly adopt Gen AI recommendations? And if a recommendation turns out to be unsuccessful or wrong, who’s to blame? The technology provider and co-creation teams? Gen AI? The employee? All parties involved? These are some things to consider when diving into the world of Gen AI.