00 Opening Name the work
Websites, apps, automation, and agentic systems

Build working AI workflows around real work.

I build websites and workflow tools for individuals and teams. I also build automations, desktop apps, mobile apps, and agentic systems. If you are not sure what to build first, I help you choose the right starting point.

StartChoose one workflow that will create meaningful improvement.
BuildOrganize source material, write the AI instructions, and build the first usable version.
CheckKeep human judgment visible before decisions are finalized.
Workbench with notes, devices, and review material for AI workflow design.
01 Problem Tools need workflow

What makes AI work valuable

Most teams already have AI tools. The harder job is choosing the workflow and organizing the source material. The work also needs working memory, a named reviewer, and proof that the output is ready to use.

01

Pick the workflow.

Name the task, owner, inputs, and output people rely on.

02

Clearly define what the AI is allowed to do.

Decide which tasks AI can draft or summarize. It may also compare, prepare, or route work.

03

Review before use.

Define the human review, source verification, risk review, and handoff process before the work moves forward.

02 Services Where to start

Where to start

Start with the path that matches the current need. Some clients know what they want built. Some need help choosing the workflow first. Some need advisory support while their team builds.

Build

Build It For Me

Have me build the website or app you need. Workflow tools, automations, and agentic systems fit here too.

Review Services

Starter

AI Workflow Starter

Turn a task you handle often into a working AI-assisted workflow with source data, AI instructions, and a review step.

Review Services

Advise

Build Advisory

Get help with scope and source setup while you or your team builds. Use it for agent behavior, review checks, and implementation choices too.

Scope a Build

03 Capability What changes

The capability ladder

AI access becomes useful when the work is shaped into repeatable capability. The ladder shows what has to become clearer as the work matures.

01

Access

People are using AI, but the process is still inconsistent from case to case.

02

Usage

People use AI for drafts, notes, summaries, and first-pass thinking.

03

Delegation

The team decides which tasks AI can support and which decisions stay human-owned.

04

Verification

Outputs are checked against source material, risk, usefulness, and by the person who owns the result.

05

Redesign

The workflow changes so source material, review, and output are easier to repeat.

06

Operating system

The work leaves behind reusable habits, records, tools, and review paths.

04 Proof Inspect the method

Proof you can inspect

Sample audit

A sample audit output for a workflow worth improving.

See the workflow, source material, and rules for what AI can do. The sample also shows the review checklist, working first version, and first implementation path.

Inspect the sample audit

Workflow diagram

A bounded agentic workflow from trigger to review.

The diagram shows the input, source material, and AI role. It also shows tool access, human review, and the next action.

See the workflow diagram

Pathway Protocol

A working example of AI task routing and review.

The protocol shows Think path, Build path, and Safe Lane. It also shows handoff rules, checks, and human review.

See the teardown

Audit framework

A practical way to choose the first workflow.

The audit framework shows how to map current work, name what AI can support, and decide what stays human-owned.

Read the framework

05 Writing Public proof layer

Writing that explains the method

Writing supports the buyer path. It shows the method behind the audit, the build work, and the review standard.

The AI Capability Audit

A framework for choosing the first workflow to improve.

Read it

The Multi-Agent Orchestration Protocol

A practical method for task routing, review, and handoff notes.

Read it

Just Enough Intelligence

A working note on matching model effort to the task.

Read it