Evaluate the whole assignment before accepting a persuasive score.
EXECUTIVE ANALYSIS | H. Max Ottesen
AI-generated conceptual editorial illustration. Fictional people and setting; not documentary evidence. Overhead illustrated transportation planning desk with route paper, keys and maintenance information.
Would you accept a load because its quoted rate looks attractive? I would want the rest of the assignment beside it: empty movement, waiting, equipment fit, available driver time, customer obligations and the next positioning choice. That is the decision I would ask a transportation tool to help prepare.
The proposed application in this article is an advisory comparison of operating options. It is not a description of a tested optimizer or autonomous dispatcher.
Show the tradeoff
A useful comparison should let a dispatcher inspect the assumptions for each option. Which distance is being used? Is the receiving window confirmed? Is the equipment appropriate? What cost is estimated, and which cost comes from an actual record?
I would keep safety and operating eligibility outside a negotiable commercial score. An attractive revenue estimate should not compensate mathematically for an unresolved maintenance hold. The review should first reject options that fail the applicable operating checks, then compare the eligible choices.
EPA’s SmartWay resources combine performance measurement with operating and equipment strategies. [1] They support a broad view of fleet efficiency, not a claim that AI creates a particular saving. My suggested tool would similarly show several operating considerations without pretending they reduce to a universally correct ranking.
Keep the career example in its proper place
My Career Vault records on-time performance moving from 72% to 94% and transportation cost per mile from USD 3.75 to USD 3.02 at PetroChoice / MOOVE. The figures are checked against my approved career record; independent employer calculations were not supplied. They are not AI results. [2]
I would use that experience to ask better questions of a new tool. Are cost and service being measured on comparable work? Did the customer mix change? Are savings shifted to another budget? These are proposed review questions, not explanations I can establish from the Vault alone.
Compare decisions before automating them
Select one repeated choice, such as late-load recovery. Record the existing dispatcher recommendation, the tool’s proposed alternative and the information available at the time. Avoid judging a recommendation using facts that arrived only after the decision.
Begin with retrospective cases, then a supervised trial. Track why dispatchers override suggestions. An override may reveal a missing constraint, but it may also reflect habit. Both deserve examination. Include the time spent correcting data and reviewing recommendations in the evaluation.
Research with consultants described by Harvard Business School found that AI capability varied across tasks. [3] That is a reason to test each intended transportation use. It is not evidence that a model performing well on consulting exercises can choose loads safely or profitably.
The pilot should earn further authority only if it helps people make better decisions under the actual constraints. A neat score is easy to display. An assumption that a dispatcher can inspect, challenge and correct is more useful.
SOURCES & EVIDENCE
[1] U.S. Environmental Protection Agency (Undated live page; accessed 2026-10-17). SmartWay Truck Carrier Partner Resources. https://www.epa.gov/smartway/smartway-truck-carrier-partner-resources. Locator: Measure your Performance and Strategies for Efficiency. Supports: Fleet measurement and multiple efficiency strategies. Limitations: Describes measurement and operating strategies, not AI-caused savings or Max’s fleet results. Checked: 2026-10-17.
[2] H. Max Ottesen (2026-10-15). Executive Career Vault and Resume Rules. CAREER_VAULT_SOURCE.pdf, Version 3 Final, sections 5–7, pp. 2–4. Locator: CAREER_VAULT_SOURCE.pdf, Version 3 Final, sections 5–7, pp. 2–4. Supports: The two historical PetroChoice / MOOVE measures, without AI attribution. Limitations: Verified against the governing Vault only. Employer records, underlying calculation definitions and independent audit evidence were not supplied. Checked: 2026-10-17.
[3] HBS AI Faculty / Harvard Business School (2023-09-21). Navigating the Jagged Technological Frontier. https://aiinstitute.hbs.edu/navigating-the-jagged-technological-frontier/. Locator: Study overview and Key Findings. Supports: Task-specific capability in a consultant experiment. Limitations: Study involved 758 consultants and task-specific performance. It does not establish fleet, warehouse, safety or general software-production outcomes. Checked: 2026-10-17.
HOW THE CONCLUSION WAS REACHED
The operating record provides context, EPA supports multi-factor measurement, and the research cautions against assuming capability transfers between tasks. The proposed decision comparison follows from those premises but remains an untested design.
LIMITATIONS
No load-level cost data, optimizer evaluation or causal savings analysis was supplied. Historical measures lack underlying definitions and independent corroboration. Consultant research is not transportation evidence.
© 2026 H. Max Ottesen. All rights reserved.
← All articles & insights