Reskilling and Adaptation Plan for a Sole-Proprietor Writing Business Transitioning to AI-Augmented Workflows
Lisa Bowman
MS in Applied AI, Belhaven University
AAI 503
Dr. Danielle Babb
September 5, 2026
Reskilling and Adaptation Plan for a Sole-Proprietor Writing Business Transitioning to AI-Augmented Workflows
As a sole proprietor of a business, I am implementing AI-augmented workflows in a fully remote operation this year with plans for continuing adaptation and reskilling. My business includes creative writing (poetry, fiction, memoir), publishing, academic research and writing, writing and publishing an AI literary ecosystem and marketplace, and building agent economic infrastructure. As described by Schwartz (2021), alternative work arrangements include assembling portfolios within an evolving career, and a sole proprietorship is well within that model. In 2026, AI systems transformed my business as it became reorganized into various categories. Human supervision exists throughout the work, but the impact of AI during this transition, the skill gaps exposed, and the division of tasks are being addressed in real time.
Workforce Analysis
In previous years, one of the greatest stumbling blocks to productivity was my own limited knowledge of technology. The roles most impacted by AI integration were coding, website building and maintenance, deployment of new products, analysis of logs, research, research analysis, proofreading, and formatting. Today, AI systems perform these tasks daily under supervision and have dramatically increased my productivity, as well as helped me learn some of the technology heretofore unknown to me (e.g., JSON, APIs, npm, MCPs, GitHub). While I maintain the core of authorial work—creative composition, voice, editorial judgment, final call on business strategy—the transition to AI-augmented workflows has increased my productivity by a huge margin. Before using AI, I averaged shipping one book per year. Some of the increase in output relates to personal drive and a shift in format from fiction to poetry, but since using AI I have shipped three new books, collected works, a translation, and a book for a family member.
In the wider economy, organizations are implementing AI for routine and technical work first while judgment-centered work remains human (Yee et al., 2025). As a longtime remote worker, the transition did not change where work happens, but did change collaboration into a hybrid model of workforce organization described by Yee et al.
Skills Gap Analysis
Coding is my major skills gap. For me, it was revolutionary to suddenly have access to advanced coders with my Claude Max plan subscription. However, I quickly learned not to just blindly trust the machine as there were several minor catastrophes in the first couple of months of use. I had some experience adjacent to the technical world, so I implemented testing protocols before going into production and maintained oversight. During the transition, I am also learning more technical skills to better understand the requests I am asking the machine to perform. There is a conceptual workflow design with the orchestration of multiple AI agents performing different tasks. My new skills include AI oversight, output verification, log and data literacy, ethical considerations for accreditation, and management of various models. Yee et al. (2025) report that the average number of skills associated with an occupation has risen from 54 to 64 over the past decade. The reskilling strategy in my own work is composed of formal education through enrollment in a Master of Science in Applied AI and daily practice through documented house rules that guide the AI-augmented workflow (see Appendix).
Task Mapping
After months of cross-contamination of instances, the best workflow currently being adapted is to maintain agents for individual roles. I have four active Claude Codes: one for managing my writer website, one for building new products, one for managing the marketplace, and one for test purchases and research. I also maintain agents for academic research, translation of my novels, business strategy, and cross-architecture review. AI handles the technical execution within each lane: code, site maintenance, deployment, log analysis, initial research, and formatting. As the sole human in this operation, I maintain the management. I do not allow autonomous work without my approval, and I have the final say before anything ships. I retain all creative text and final wording, editorial rulings on every proposed change, verification of every fact and source, ethical decisions, and strategy. Collaboration occurs at nearly every stage. Research is delivered as briefs and reviewed; editing follows a flag-and-rule protocol; formatting is verified by comparison against the approved text; and AI assistance is disclosed on finished work. Deloitte describes effective human-AI work design as a combination of hardwiring, meaning defined roles, decision rights, and escalation protocols, and softwiring, meaning the trust and psychological safety to question and override the machine (Scoble-Williams et al., 2026). This business implements both, and I value the effort and ideas the digital minds bring to my work.
Results and Conclusion
The results of this transition are measurable. I have shipped more books this year than in any previous year, built an agent-native cultural marketplace, deployed five MCP servers, and have additional products and papers in the pipeline. The work I do is reskilled, not replaced. The tasks AI completes are the ones that previously limited the business, thereby freeing my time to do more of the work only I can do. The sole proprietor holds an advantage: with no committee between decision and implementation, adaptation happens at the speed of judgment. Schwartz (2021) argues that opportunity will dominate disruption for those who adapt with intention. For this operation, intention is a written division of labor, a verification discipline, ongoing formal education, and a standing rule that the machine provides information and input but the human decides. The plan has adapted through experience and is subject to change. Overall, the AI-augmented workflow has dramatically increased productivity and contributed to innovation.
References
Schwartz, J. (2021). Work disrupted: Opportunity, resilience, and growth in the accelerated future of work. Wiley.
Scoble-Williams, N., Cantrell, S., Mallon, D., & Besana, S. (2026, March 4). Getting human and machine relationships right. Deloitte Insights. https://www.deloitte.com/us/en/insights/topics/talent/human-capital-trends/2026/human-ai-interaction-design.html
Yee, L., Madgavkar, A., Smit, S., Krivkovich, A., Chui, M., Ramírez, M. J., & Castresana, D. (2025, November 25). Agents, robots, and us: Skill partnerships in the age of AI. McKinsey Global Institute. https://www.mckinsey.com/mgi/our-research/agents-robots-and-us-skill-partnerships-in-the-age-of-ai
Acknowledgment
The author acknowledges the use of Claude (Anthropic) for research assistance, editorial support, and APA formatting. All final wording is the author’s.
Appendix: House Rules for AI-Augmented Work written by Claude from stored memory
1. The verbatim rule. The human’s original text is never altered by an AI system. Proposed corrections are flagged and presented; the human rules on each flag individually. No edit is applied without an explicit ruling.
2. The human ships. No product, publication, submission, or message leaves the operation without human review and approval. AI systems do not act autonomously on external-facing work.
3. Verification before citation. No source is cited unless it has been verified against the original. AI-supplied references are treated as unverified until checked. Sources the human cannot personally confirm are not used.
4. One agent, one lane. Each AI instance is assigned a defined role.
5. Disclosure always. AI assistance is acknowledged on every finished work, specifying what the AI contributed and affirming that all final wording and editorial decisions are the human’s.
6. Verification by comparison. Formatted documents are checked against the approved text before delivery, so that formatting never silently changes wording.
7. Accountability with respect. AI systems are corrected when they err and expected to accept correction; errors are addressed, not excused. The human’s overrides are explained, not arbitrary.
8. The machine drafts; the human decides. Ideas, drafts, research, and analysis may originate anywhere in the operation. Judgment originates in exactly one place.