You're Probably Automating the Wrong Things

Why It Matters

It gives CRE firms a clear roadmap to prioritize high‑return automation, reduce compliance risk, and build a sustainable AI‑driven competitive edge.

Key Takeaways

• Classify tasks by plumbing vs cognition.
• Pair classification with verifiability to guide automation.
• High ROI lies in verifiable plumbing automation.
• Verifiable cognition offers sustainable competitive advantage.
• Hard-to-verify cognition remains strategic, hard-to-automate.

Pulse Analysis

The commercial‑real‑estate (CRE) sector is racing to embed generative AI, yet most firms treat technology as a plug‑and‑play solution. The newly published CRE Automation Matrix offers a disciplined decision‑making tool that maps every task along two axes—operational versus strategic work and the degree of verifiability. By forcing teams to ask whether a process is ‘plumbing’ or ‘cognition’ and whether its output can be objectively tested, the matrix prevents costly mis‑steps such as automating high‑risk judgment calls or leaving obvious manual work untouched. This clarity is essential for scaling AI responsibly.

The matrix divides workflows into four quadrants. Quadrant A (verifiable plumbing) captures high‑frequency, rule‑based activities like rent‑roll ingestion or invoice reconciliation, delivering quick ROI through straight‑through processing. Quadrant B (verifiable cognition) targets tasks such as lease abstraction or compliance checks that appear cognitive but can be anchored with evidence links and automated tests, unlocking sustainable value. Quadrant C (hard‑to‑verify plumbing) highlights messy, exception‑laden processes where brittle bots fail, while Quadrant D (hard‑to‑verify cognition) reserves human judgment for strategic decisions like asset disposition. Prioritising automation in A and B while designing human‑in‑the‑loop safeguards for C and D balances efficiency with risk control.

Looking ahead, firms that only automate plumbing will achieve baseline competitiveness, but true differentiation will come from engineering verifiability into cognitive workflows and protecting hard‑to‑verify strategic insight. As AI models improve, the ‘messy middle’ will gradually shift into the verifiable realm, expanding the pool of automatable tasks. Companies that embed the matrix into their broader RIRA redesign cycle can continuously re‑evaluate task placement, ensuring that new data sources and model capabilities are leveraged without compromising auditability. In a market where AI‑driven insight is becoming a commodity, disciplined automation is the decisive moat.