On document extraction and citation traceability -- the tasks that underpin the entire workflow -- Tower performs above the category average and within 6 points of the GPT-5.4 frontier on the Lab's structured extraction task battery. The product's structural edge is the combination of extraction capability with workflow infrastructure: automated data room organisation, document naming, missing document flagging, and role-based access controls in a single platform. This combination is what general-purpose LLMs cannot replicate without significant integration overhead.

The frontier is improving at 3.6 points per quarter on document extraction tasks. Tower's extraction gap from GPT-5.4 is 5.8 points -- manageable but compressing. The product's durable advantage is the workflow layer: data collection, request list management, missing document detection, team collaboration, and access controls are features that a frontier model accessed via API cannot provide without Tower-equivalent infrastructure. The extraction engine is the commodity risk. The workflow is the moat.

For M&A lawyers and asset management teams, the buy decision for Tower is not primarily about extraction accuracy -- it is about whether replacing disconnected VDRs, Excel trackers, and email threads with a unified platform saves enough time to justify the switch. The Lab's workflow benchmarks suggest it does: automated data room organisation alone reduces the setup time in the benchmark test by an estimated 60--70% versus manual organisation. The AmLaw 50 and Chambers Band 1 deployment signal is the strongest commercial validation in the product's current category. For teams with high-volume transaction pipelines, the workflow infrastructure case is strong regardless of extraction accuracy relative to the frontier.