Will AI Kill Robotic Process Automation? - Forbes

TEXT ANALYSIS: "Will AI Kill Robotic Process Automation?"

THE DISSECTION

This is a vendor-amplified case study dressed as journalism. It uses Lemvigh-Müller as a trophy to perform the narrative that LLM-based AI is simply a better tool for automating knowledge work. The article presents the displacement of RPA as a technology upgrade story—a better mousetrap—while the actual event it documents is the systematic displacement of cognitive labor. The framing sanitizes structural collapse into an efficiency victory lap.

THE CORE FALLACY

The article treats the transition from RPA to LLM-based AI as a lateral tool substitution: one automation technology replacing another. It is not. This is a category-ending event.

RPA automated rule-following at the GUI level. It required structured, repetitive inputs and could not handle variability. LLM-based AI reads unstructured PDFs, performs cross-document comparison, identifies deviations, and drafts responsive communications—all cognitive tasks previously requiring human judgment. The article celebrates this as "AI reading PDFs," but the mechanical reality is that the 5,000–7,000 man-hours being eliminated per year are not the hours of robots clicking. They are the hours of knowledge workers performing cognitive comparison and decision tasks. The humans are not being elevated. They are being removed from the circuit.

The article implicitly treats this as a win for the economy. Under DT mechanics, this is the precise mechanism of productive participation collapse: cognitive work being extracted from the wage circuit and performed by capital (the LLM infrastructure) at near-zero marginal cost.

HIDDEN ASSUMPTIONS

1. Human workers are interchangeable with software — The article never addresses what happens to the workers whose hours are eliminated. They are treated as inputs, not stakeholders. The assumption is that their displacement is simply a cost reduction event.

2. 98% accuracy is an acceptable threshold — The article presents this as a success metric without examining what the remaining 2% represents at scale, or who bears the risk when it fails.

3. Cost per document is the only relevant metric — The framing is purely transactional. The systemic cost—the elimination of an entire category of employable cognitive labor—is treated as external to the analysis.

4. SAP's "safety" and dashboard are meaningful moats — These are integration layer features, not defensible capabilities. They add cost and vendor lock-in without addressing the underlying displacement.

5. "Transition to AI in other areas" is a neutral expansion — The article notes the company will apply the same model to invoices, delivery date PDFs, and inbound orders. This is not a case study. It is a deployment blueprint for the systematic elimination of back-office cognitive employment across the supply chain.

SOCIAL FUNCTION

This article performs transition management propaganda. Its function is to normalize the displacement of cognitive workers by framing it as an efficiency improvement, a tool upgrade, and a competitive advantage. It celebrates the human who suggested using ChatGPT as if this is empowerment, when in fact the human's suggestion led to the elimination of their colleagues' employment. The article is SAP marketing infrastructure operating under a Forbes byline, and it is entirely useful to the Sovereign class deploying this technology: it provides plausible deniability that the human cost is being considered, while documenting the playbook for accelerating displacement.

THE VERDICT

This article is a deployment memo for productive participation collapse, published under the editorial license of a technology success story. The 5,000–7,000 eliminated man-hours at one Danish wholesaler are not an anomaly. They are a proof-of-concept for every PDF, every invoice, every order confirmation, every structured document in every supply chain on Earth. Lemvigh-Müller is not an outlier. It is the first layer of what will become a geological event: the systematic removal of cognitive labor from the wage circuit at a scale that has no historical precedent.

RPA is not being killed. It is being buried. And the grave is being dug by the workers it will replace.