Start with work I actually did.
Real decisions, constraints, observations and lessons—not a generic summary anyone could generate.
First-hand operating notes
Build what is useful. Show what is true.
Field notes from designing practical AI agents, websites, search systems, CRM workflows and lead paths. No recycled “AI hacks,” invented results or ranking guarantees—just the decisions, controls and lessons that hold up in real work.
Latest field notes
Original experience, clearly documented.
What I am building and learning now.
How does AI-assisted sales work for contractors?
AI-assisted contractor sales organizes incoming project information, prepares responses and flags missing next steps while a person owns customer decisions. Start with one verified inquiry path: capture the request, assign an owner, review the scope and schedule the next action. Confirm storage and delivery before adding automated follow-up or buying more leads.
Read the guideWhat should construction company owners post on LinkedIn?
Construction company owners should post useful answers backed by work they can document: a project decision, scope tradeoff, planning checklist or lesson from a real job. Explain the question, the evidence and the next step. Link to a relevant service or guide, measure qualified conversations and obtain permission before sharing customer details or photos.
Read the guideWhat is FieldKnown?
The customer-path company Tony founded for owner-led service businesses, connecting discovery and intake to ownership, follow-up and measurement.
Read the FieldKnown definitionAI agents for small business: my 30-day human-controlled pilot.
A vendor-neutral implementation lab note: choose one workflow, establish a baseline, restrict the tools, test failure paths and keep every external action behind Tony’s approval.
Read the 30-day pilotJeff: a useful AI agent with a human decision at the edge.
One bounded job, approved information, narrow permissions, source-backed evidence and a person deciding what happens next.
See Jeff’s controlsHow I build useful and controlled AI agents.
The operating model behind the pilot: one orchestrator, narrow tools, evidence, stop rules and measurable business outcomes.
Read the operating modelWhat is revenue operations?
A plain-English operating model connecting marketing, sales and customer follow-through through shared stages, ownership, data and next actions.
Read the RevOps guideHow AI supports a revenue pipeline without owning it.
A human-controlled pipeline architecture for market mapping, account evidence, CRM ownership, commitments, follow-up and measurement.
Inspect the pipeline modelAI-assisted outreach starts with evidence—not volume.
A responsible research and drafting workflow built around narrow targeting, public evidence, human approval and CRM follow-through.
Read the outreach workflowAll-Pro: a documented local customer path.
A first-party implementation record for service, location, proof and inquiry paths. Commercial outcomes remain awaiting a dated, source-system baseline.
Inspect the All-Pro implementationA human-controlled B2B AI outreach system.
A practical ICP, account-research, message-preparation and CRM workflow that keeps every external send under human approval.
Inspect the responsible system Editorial standard
Built for people. Understandable by machines.
Three rules for every field note.
Separate proof from aspiration.
Sources, dates and inspectable facts are named. Unknown results stay unknown until measurement supports them.
Give the reader a working model.
Each note explains what to do, what to avoid and where a responsible human must remain involved.
Bring a real workflow
What is the business system you wish worked better?
Share the goal, the tools already in place and where the handoff breaks. I will identify the first system worth reviewing.
From reading to working.
Bring Tony one bottleneck.
Start with the handoff, repeated task, search path, lead process or CRM gap that is costing the business attention.
See ways to work with Tony