Greg Holmes, Field CTO for EMEA at Apptio (an IBM company), argues that successfully scaling intelligent automation requires financial rigor.

Unit Economics and Scalability

Many innovation projects fail (around 80%) because the lack of financial transparency during the pilot phase hides potential future liabilities. A pilot project that demonstrates a saving of 100 hours per month may seem successful, but often does not consider that the infrastructure used is oversized compared to the real needs of a production environment.

Moving to a production environment involves an increase in computing, storage, and data transfer requirements. The number of API calls can multiply, and exceptions not considered in the pilot phase can emerge, increasing support costs.

To avoid surprises, it is essential to monitor the marginal cost on a large scale, analyzing parameters such as the cost per customer served or the cost per transaction. A business model is flawed if the cost per customer increases with the growth of the customer base.

Governance and Financial Accountability

Financial accountability should not fall solely on the finance department. Holmes suggests putting governance back in the hands of developers, integrating tools like HashiCorp Terraform and GitHub to enforce policies during deployment and estimate costs in real time.

TBM and Common Language

Often, the CFO focuses on return on investment, while the automation manager monitors operational metrics such as hours saved. The TBM (Technology Business Management) framework and Apptio aim to solve this challenge, providing a common language between technology, finance, and business.

The TBM taxonomy standardizes the reconciliation of these perspectives, mapping technical resources (computing, storage, labor) into IT towers and, subsequently, into business capabilities. This structure translates technical inputs into business outputs, providing a detailed view of service costs.

Technical Debt and Long-Term Budget

Companies with legacy ERP systems must choose between automation as a temporary solution or as a bridge to modernization. Using automation to mask inefficient processes only leads to accumulating further technical debt.

A Total Cost of Ownership (TCO) approach helps define the correct strategy. Calculating the TCO, including the hidden costs of infrastructure, labor, and engineering time, can reveal that the real cost of keeping a legacy system alive is higher than perceived, especially considering the necessary automation layers.

Balancing variable costs with long-term commitments is essential. Committing to specific technologies or platforms over a multi-year horizon allows negotiating economies of scale and standardizing the architecture, facilitating the construction of solutions suitable for the long term.