During my years as a student and later in the doctoral program, we never treated research integrity as a checklist to complete and set aside. It involved a clear set of steps I had to complete before I could include any claim in a course paper or the dissertation itself. Those same steps continue to govern the materials we produce for daily operations at Alaska Dog Works and Peak Experience, as well as the work we deliver to clients.
The Four Steps of Research Integrity
- Open the primary source. A citation in a secondary article is not enough. Locate and read the original paper, dataset, or document. If the claim does not appear there as stated, exclude it (Poff, 2024).
- Record the limitation of the source. Every source is bounded by sample, context, method, or date. State that boundary when the source is used. Findings from one industry or generation do not automatically transfer (Zhaksylyk et al., 2023).
- Treat the full text as the evidence. An abstract, review article, or paraphrase is only a pointer. Examine the complete source. Record contradictory results in the literature rather than setting them aside (Armond et al. 2024).
- Keep the chain of evidence intact. Each claim in the finished page must be traceable to the source that supports it. If you cannot show the chain, remove or rewrite the claim (Mousa et al., 2024).
These steps were required under examination. While working on the dissertation, opening the primary source meant returning repeatedly to the literature on family-firm succession to confirm exactly what each study reported before that material entered the analysis. Limitations were recorded so that a finding drawn from one industry or one generation was not presented as automatically transferable.
The full text was treated as evidence rather than an abstract or review, and contradictory results within the same body of literature were noted rather than set aside. The chain of evidence was kept intact so that each claim in the finished document could be traced and defended.
Application in Company and Client Work
At Alaska Dog Works and Peak Experience, materials prepared for internal operations and for client deliverables follow the same sequence. A source is opened before a statement is adopted. We state limitations when the available evidence does not cover the case. Any summaries and automated drafts are checked against the primary record or against what has been observed in the field. We retain the chain of evidence so a client proposal or set of curriculum pages can be defended if questioned. The process takes longer than accepting the first available text, and that extra time separates a document that holds up under scrutiny from one that doesn’t.
The same order is followed when we prepare materials for others. We check succession materials developed for family firms against the relevant literature and the practical constraints multi-generational businesses face. Leadership training pages are built from sources that can be named and from experience that can be described. The scope of the evidence is stated clearly, and verification is not replaced by the volume of material produced.
Ethical Use of AI
Artificial intelligence tools can accelerate an initial pass through a large set of sources or help produce a rough outline. We have developed an internal agent, TOGO, trained on our own materials and used for certain tasks. The agent does not replace the verification steps that follow.
Every claim drawn from an automated draft is checked against the original document or the field record before it is used. Sources are opened and read. Record limitations when the tool cannot provide them. Output that cannot be verified is discarded. The person who signs the final page remains accountable for its content.
The same standard applies to client work. Materials delivered to a family firm or used in a training session are read and revised by the people who will stand behind them. Unexamined generated text is not presented as finished work, and the verification steps are not omitted because a deadline is short. The tool remains in the role of assistant. Judgment, attribution, and responsibility stay with the people who deliver the work.
Conclusion
The standard that began in the library and under examination continues in the daily work of the companies and in the materials prepared for clients. Verification is completed before a claim is used, limitations are stated, and the chain of evidence is kept intact. Artificial intelligence may help with early drafts or source review, but it does not replace those steps or the accountability of the people who sign the final page. Clients receive work that has been checked and can be defended. You will not get AI slop when working with us.
Sources
Armond, A. C. V., Cobey, K. D., & Moher, D. (2024). Research Integrity definitions and challenges. Journal of Clinical Epidemiology, 171, 111367.
Mousa, A., Flanagan, M., Tay, C. T., Norman, R. J., Costello, M., Li, W., … & Mol, B. W. (2024). Research Integrity in Guidelines and evIDence synthesis (RIGID): a framework for assessing research integrity in guideline development and evidence synthesis. EClinicalMedicine, 74.
Poff, D. (2024). Research integrity. In Encyclopedia of Quality of Life and Well-Being Research (pp. 5965-5967). Cham: Springer International Publishing.
Zhaksylyk, A., Zimba, O., Yessirkepov, M., & Kocyigit, B. F. (2023). Research integrity: where we are and where we are heading. Journal of Korean medical science, 38(47).
