Generative AI: start small, think big

The incredibly fast adoption of ChatGPT has enlightened businesses and consumers about the recent developments of Generative AI and the endless opportunities it creates.

Less than a year after its release, most consumers are already using some form of Generative AI and 73% of them say they trust the content it generates (source: CRI report – https://www.capgemini.com/insights/research-library/creative-and-generative-ai/). Faced with this sudden worldwide enthusiasm and the promises of increased efficiency, organizations have no choice but to experiment and embrace as part of their business operations, and 49% of telco executives state that they have already started working on pilots or proof of concepts (source: Generative AI in organizations CRI report.

Generative AI can radically improve customer experience

Among corporations, CSPs (communications service providers) have been quick to grasp this opportunity to improve operations and enhance customer experiences. Indeed, managing customer interactions is complex, full of repetitive and labor- and cost-intensive text- and voice-based tasks that do not always result in high customer satisfaction. Generative AI promises to increase both efficiency and quality, thanks to agent augmentation and process automation based on wider and deeper context, as well as more relevant personalization combined with 24/7 availability. As an illustration, a recent study on the impact of Generative AI on a specific team of customer support agents identified a 14% productivity boost on average (measured by number of issues resolved per hour), with the biggest benefits observed on the novice and low-skilled workers. The study also found that Generative AI assistance improved customer sentiment and overall employee satisfaction (source).

Experiment safely with live prototyping

 Many CSPs have already started experimenting despite concerns about the maturity and the risks of the technology. If you’ve spent anytime testing ChatGPT, you’ve probably experienced the best and the worst it can offer – excellent content in some cases and totally fabricated, if convincing, drivel in others. It’s no surprise then that our clients are starting with lower-risk applications, typically opting for employee augmentation before client-facing automation. Lower risk areas are typically tasks with a higher fault tolerance, where Generative AI can offer a higher degree of automation while keeping the human in the loop and avoiding the risk of embarrassment.

Case study: assisted content creation with Generative AI

For instance, at a major telecom operator, Capgemini has developed an internal application for generating on-demand content, such as marketing product descriptions. The application uses cloud-based Generative AI APIs (application programmable interfaces), adding task and organization context via prompt engineering. In this way, Generative AI is customized for the needs of users without them having to become experts in prompt engineering. The application allows marketing teams to save time while generating more impactful and consistent content – it automatically produces high-quality product briefs, marketing content, and campaign pitches while respecting brand identity and tone of voice and complying with security requirements. The application provides a safe experimentation space because a human will always review generated content before approving it for publication.

Start small, think big

We have developed several of these low-risk applications with our clients, and CSPs should certainly seize opportunities quickly in order to gain efficiencies and familiarize the organization while the technology matures. At the same time, although these low-hanging fruits will bring efficiency gains in the short term, they will not provide a competitive advantage because everyone else will be doing the same thing. A long-lasting competitive edge will come only from a deeper and larger-scale transformation, including full context-aware automation in most customer interactions. It will require deeper understanding of how to scale Generative AI, how to make it interact with the enterprise’s data, how to make it secure, compliant, and ethical. That’s why CSPs should start building – today – for what Generative AI can offer tomorrow – more automated and personalized customer interactions and improved assisted employee experiences.

TelcoInsights is a series of posts about the latest trends and opportunities in the telecommunications industry – powered by a community of global industry experts and thought leaders.