Technology Assessment / Workflow automation · Field operations
From field notes to a useful next step.
This work brings technology consulting into a repeatable workflow: capture what is actually present, interpret the evidence, and give the operator a plan they can review and explain.
Illustrative example
- 01 / ObserveBackup status is unknown.
- 02 / InterpretRecord an evidence gap.
- 03 / Follow upVerify before approving the plan.
The problem
An office assessment gathers uneven information about devices, backup, software, and infrastructure. Turning those observations into a consistent plan takes judgment. Missing information must stay visible, and related findings should not create duplicate work or costs.
My contribution
I translated experience from my technology consulting practice into assessment criteria, report logic, and workflow prototypes. I connected the earlier automation across WordPress, Make, Airtable, and Claude, then developed the current application with AI-assisted development around explicit records, deterministic evaluation, and human review.
Follow the workflow
Earlier workflow: connect the tools
A WordPress assessment form fed Make and Airtable. Formulas structured report text and tiered pricing, and Claude helped generate an executive PDF debrief. This was an internal integration prototype.
Current application: preserve the observations
The Next.js application stores clients, sites, assessments, assets, and observations in SQLite. Evidence belongs to a record rather than only to an automation run.
Evaluate before writing the narrative
Deterministic rules distinguish strengths, response items, missing evidence, and optional upgrades. Related responses are grouped for sequencing and cost calculations.
Review the proposed plan
Report generation and approval versions preserve the review process. An optional AI-written executive debrief has a deterministic fallback; the operator still checks the findings against field notes.
Decisions and tradeoffs
Unknown is a useful answer
An unverified backup becomes an evidence gap, not a confident claim that backup is present or absent. This preserves a concrete task for the assessor to follow up.
Rules own the evaluation
Priorities, finding categories, and grouped costs come from explicit application logic. AI can assist the written debrief without being the sole source of the assessment’s conclusions.
Make the workflow inspectable
Moving from a chain of services to versioned application code and persisted records makes state easier to inspect. It also introduces responsibility for access, migrations, recovery, and completing the operator journey.
Explore the evidence
An implemented evaluation model
Existing fixture checks cover an all-positive device assessment, mixed findings with missing evidence and grouped costs, and an absent optional asset. They provide focused checks of the rules rather than proof of the complete service.
A distinct application generation
The current application implements SQLite persistence, deterministic findings, report generation, and approval records. These are separate from the earlier WordPress-to-PDF integration.
A concrete consulting use case
The public assessment page explains the service and provides a path to discuss a real office-technology problem.
Read about the assessment serviceWhere it stands
The current application remains an internal workflow in development. Resuming drafts and producing a real export from an approved report still need completion and end-to-end verification. It is not a public client portal. The earlier PDF workflow should not be mistaken for an export feature in the current application.
The next useful improvement
Complete one synthetic assessment from draft through review, approved report export, and recovery after a restart. Keep the public service explanation separate from the operator application.