A new approach to CX analytics, using simulation + generative AI
Skip to content

A new approach to CX analytics, using simulation + generative AI

The experience that customers have around products and services (CX) can be a critical factor in the success of the companies providing them.

Voices of Tomorrow is a Decision Intelligence platform that makes it easy for leaders to anticipate the impact of actions on customer experience.


The use of simulations within business decision making is a $20bn industry. But the tools used to build, run and analyse these simulations are clunky, complex and hard to use - particularly for non-technical teams. If insights are not understood by business leaders, they won’t be acted on. If they are not acted on, value is not realised. Our approach is to completely rethink the user experience of building and understanding predictive simulations so that decisions are made, actions are taken and outcomes are actually achieved.

By combining systems dynamics and agent-based simulations, with natural language generation and intuitive data visualisation, it will be the world’s first CX analytics platform to deliver predictive insights in a way that everyone can understand - natural human language.


The platform provides a no-code interface that allows users to configure simulations that not only infer the likely outcomes of different decisions, but make it easy to audit the assumptions, weights and biases behind them. This is particularly important in a time when decisions are increasingly made algorithmically that have a material impact on people’s lives.


Voices of Tomorrow not only makes it easier for firms to gain foresight into how decisions will impact niche segments of people but it will also simplify the process of auditing those decisions to understand the contemporaneous expectations of impact on different groups.


Users - typically consultants working with client-side subject matter experts - build causal decision graphs for their use case via a drag and drop ‘patch interface’ in a browser-based web app. In doing this, they are defining a model based on the known / modelled / assumed causal dependencies between various factors and will define relevant demographic details for the subjects of the simulation.

Once the simulation is run, the likelihood of various outcomes is computed. Scenarios can be sampled from the distribution of outcomes. In addition to visualising these outputs graphically within the application’s interface, Voices of Tomorrow also converts these numerical outputs into natural language – first as prompts for large language models, (pre-trained for different use cases with context-specific historical "Voice of the Customer" data) and then, by the LLM, into first person statements or ‘scripts’.


Based on the demographic information defined by users, photorealistic human faces of archetypal ‘customers’ are generated. These faces are then used alongside the scripts generated by the LLM as inputs for life-like videos to be rendered at runtime, using another generative AI.


Like what you see?

Voices of Tomorrow is an ongoing collaboration between Parallel and Decision Lab. Get in touch to find out more.

Let's talk


See all