How an Automotive Supplier Saves €708,000 a Year with Agentic Automation

From unstable RPA bots to enterprise-wide automation in 12 months. A case study.

When the board of a global automotive parts manufacturer approved investment in RPA at the end of 2021, the vision was clear: reduce costs, increase productivity, secure competitive advantage. What followed was less clear. The first automations were unstable. The savings too modest to justify the effort. Internal developers struggled with a technology they had never been properly trained on. And management was growing impatient.

What this client experienced is far from unique. Gartner reports that over 50 percent of all GenAI projects are abandoned after the proof of concept. The cause is rarely the technology itself. It is the way organisations deploy it: without a clear strategy, without process understanding, without a scaling framework.

This case study shows a different path. How a company with over 300 potential automation candidates moved from failed bot experiments to an enterprise-wide automation platform — saving €708,000 a year in the process.

The Starting Point: Much Invested, Little Achieved

The client — a globally active automotive parts manufacturer with shared service centres in multiple countries — had already invested in UiPath licences and built an internal RPA team. The result after several months: a handful of automations that regularly broke down, dissatisfied stakeholders in the business units, and management expecting measurable results within a year.

KEY CHALLENGES

• Over 300 processes were queued for assessment — but there was no systematic approach to identify actual automation potential.
• Internal RPA developers were insufficiently trained. The bots built were unstable, causing frustration among employees and stakeholders.
• There was no prioritisation framework: which processes deliver the highest € impact? Which are technically feasible? Which provide the fastest ROI?
• Management expected measurable results within 12 months — not another pilot project.

In short: the company had the right idea — but the wrong approach. And it was not alone. McKinsey reports that only one third of companies achieve enterprise-wide AI scaling. The rest get stuck in what the industry calls Pilot Purgatory.

The Approach: Understand First, Automate Second

The client engaged Lunatec with a clear mandate: define a strategy and process for the entire automation lifecycle — from identification to operations. No more pilots. A system that scales.

The decisive difference from the previous approach: Lunatec started with the business, not the technology. The first question was not: which process can be automated? But rather: which process has the highest impact lever?

Phase 1: Top-Down Rapid Assessment

In the first round, all processes in the shared service centre were evaluated against the service catalogue — not by technical feasibility, but by three criteria: resource impact, time impact, and quality impact. From over 300 candidates, the most promising were shortlisted.

Phase 2: Prioritisation and Detailed Assessment

Together with the business units — Finance, Controlling, Procurement, Logistics — Lunatec architects assessed the shortlisted processes in detail. Process documentation was reviewed, pain points captured, automation potential quantified. For each prioritised process, a target process design was developed: not the old workflow with a bot on top, but a rethought end-to-end that incorporated the strengths of automation from the outset.

Phase 3: Implementation and Scaling

Implementation followed a structured automation lifecycle. Lunatec provided a dedicated team that developed, tested and deployed automations to production. In parallel, the team continuously identified new use cases — an ongoing process, not a one-off project.

The Results: Numbers That Speak for Themselves

After 12 months, the picture looked exactly as management had hoped from day one — except now it was based on hard data, not hope.

51	18	41.583	708.798 EUR
Use Cases in Pipeline	Use Cases Live	Hours Saved / Year	Annual Cost Savings

Particularly noteworthy: break-even was reached as early as month seven. From that point, cumulative savings exceeded cumulative costs — and the gap widened with every subsequent month.

The investment structure was transparent and predictable, no hidden budget, no surprises.

Where the Greatest Lever Are

A look at the distribution of automation potential by department shows why top-down prioritisation was so critical. The greatest potential lay in areas that at first glance are not necessarily the most obvious candidates.

Finance alone represented a potential of almost 35,000 hours per year — of which nearly 16,000 had already been realised by the end of the first year. Procurement and Logistics followed as the second and third largest levers. In some departments — HR, IT, Electronics — the potential was identified but not yet actioned. That is not failure, but part of the plan: the pipeline grows faster than implementation capacity, which creates a healthy dynamic.

What Other Organisations Can Learn From This

This case study is not unique. But it illustrates principles that McKinsey, Deloitte and Bain consistently identify as success factors for agentic automation.

Strategy before technology. The client already had the technology — what was missing was the systematic framework. Investing in a rapid assessment of all processes before the first build was the decisive difference.

Prioritise by € impact, not technical feasibility. The simplest processes are rarely the most valuable. A top-down prioritisation by business impact rather than technical difficulty ensures the first results are also the most compelling.

Process redesign, not process copy. The most common mistake in RPA: automating the old workflow 1:1. The Lunatec approach — design the target process first, then automate — avoids the trap of automating inefficiencies.

Managed services, not build-and-forget. Automations are not software releases you ship and forget. They require monitoring, optimisation and continuous development. The managed service model ensured the 18 deployed use cases ran reliably — and that the pipeline kept growing.

Transparency as a success factor. The automated reporting structure was not a by-product but a strategic element. Management could see at any time where savings stood, which use cases were in the pipeline, and how the investment was developing. That built trust — and budget for the next phase.

Outlook: From RPA to Agentic AI

With 51 use cases in the pipeline and 18 already in production, the client has built a foundation on which the next evolutionary step can be built: Agentic AI. While the automations to date consist predominantly of rule-based RPA bots and hybrid agent solutions, the next wave of potential lies in fully autonomous, context-aware agents that do not just execute — but decide.

The foundation is in place: clean process documentation, a live pipeline, a proven operating model, and management that trusts automation — because it has seen the numbers.

The €708,000 saved in year one is not an endpoint. It is the proof of concept for everything that follows.