At Agile Networks Technologies (ANT) we have a strong track record in simplifying complex digital selling journeys and integrating them into our customers’ core systems and scalable infrastructure. But we are not just a delivery partner, we also invest in research and development to make CX strategy execution faster, measurable, and repeatable. In one word: scalable. A clear example is our in-house machine-learning consulting approach we called LEADER, built to propose and implement tailored selling journeys using a structured library of journey “building blocks” developed from real-world projects.

That research and development (R&D) focus was formally validated through a funded innovation project in Germany, for which ANT has received the Bescheinigungsstelle Forschungszulage (BSFZ) certificate that validates our R&D activity and allows us to claim funding from the research allowance. The project’s goal was to significantly reduce the time spent on documentation, project planning, and effort estimation, tasks that are traditionally done manually with general purpose software, like Word, Excel and PowerPoint, by moving toward data-driven automation. In the project, we explored and applied unsupervised learning approaches that can interpret user stories and match them to the most relevant customer journey building blocks, generating a recommended selling journey and a corresponding delivery plan.

As part of the work, we developed machine-learning algorithms that compare a new customer project against a standard selling journey, then propose a project plan and model the expected effort. This enables real-time tracking of budgets and timelines and supports proactive delivery management: teams can identify early signals of delay and decide on corrective actions based on objective indicators rather than manual updates and assumptions. The result is a measurable productivity gain in planning and governance, freeing up capacity for higher-value consulting and customer-facing work.

It is important to acknowledge that this was not routine development. The project extended beyond our established methods by combining CX performance indicators – such as experience attributes, customer loyalty and Net Promoter Score (NPS) – with statistical modelling to optimize outcomes, and by applying additional techniques like coefficient reduction, expanded datasets, and cross-validation to reach a commercially reliable solution. While we already had experience with regression-related methodologies (including factor analysis and state-space models), the funded project required us to build further capabilities in unsupervised learning and advanced statistical modelling (including Markov Chain Monte Carlo, MCMC), underlining its R&D nature and its contribution as a novel extension beyond our day-to-day delivery.