Find the leak.
Build the fix.

The challenge

Teams at a national crane company entered the same information in disconnected systems, creating conflicting records and an unreliable view of the business.

What changed

A reconciled data foundation and actionable reporting helped managers review payroll discrepancies, investigate job profitability, coach pricing, and focus inspection renewal outreach.

What I built

• A reconciled data lake with a dedicated reporting layer, keeping report queries off production systems.
• Technician time reconciliation with manager explanations and regional approval.
• Job profitability, sales, pricing governance, and inspection renewal dashboards.

Approximately $71,000 in calculated payroll savings by July 2026, measured against the pre-rollout baseline.

The challenge

A week of building surveys and travel left the company’s estimator spending weekends turning photos and notes into estimates and proposals.

What changed

The app and AI estimating assistant are in production. Submitted surveys generate estimates that can be revised by text. It’s too early to measure time saved.

What I built

A mobile survey app connected to an AI estimating assistant, an approved pricing catalog, and a standardized estimating form. The assistant automatically generates estimates from submitted surveys and revises them through text conversations. The estimator reviews and approves the pricing, stepping in only to adjust markup or price unusual requirements.

What’s next

I’m expanding the assistant to generate customer proposals, branded survey reports, and fire life safety cut sheets, and connecting document delivery—with human approval before anything goes to the customer.

The challenge

Project managers needed job details and follow-ups while working in the field. Contracts, emails, photos, and promises were scattered, leaving too much to memory.

What changed

Two project managers and one operations manager use Hal daily. They no longer have to remember every follow-up or dig through emails. Hal brings what needs attention to their phones each morning, and they update him while working in the field.

What I built

Hal, an AI project assistant that files notes and documents by job, answers from project records, and tracks commitments. His weekday morning brief organizes the day into:

• Money to follow up on — unresolved payments, quotes, and approvals.
• Do today — calls, confirmations, and next actions.
• Waiting on others — outstanding answers and decisions.

The challenge

Frequent travel made it difficult to keep up with customer emails and a changing schedule without repeatedly opening a laptop.

What changed

The estimator can manage email and calendar tasks through conversations with Ash while on the road. Ash flags customers who need immediate attention and handles follow-ups when instructed.

What I built

Ash, an AI email and calendar assistant with read and write access to the estimator’s mailbox and calendar.

• Monitors incoming email and alerts him to urgent customer needs.
• Writes in his voice and sends emails on his behalf when instructed, including follow-ups.
• Sets reminders, adds meetings, rearranges appointments, and helps plan the week.

The challenge

Company knowledge needs to be usable beyond the person who knows it or the document where it was first recorded. AI assistants need that business context, too.

What changed

The business now has a shared reference that people and AI-assisted workflows can reuse. Connected routines draw on its records to prepare briefs and updates, giving automation a foundation in how the company actually works.

What I built

A structured company knowledge repository for an elevator WBO company.

• Organized company facts, procedures, technical references, and reusable templates.
• Recorded sources, review dates, and approval status so current guidance is distinguishable from drafts.
• Created a foundation for AI assistants to find relevant information and support automations, with access defined for each workflow.

The challenge

Keeping up with industry news, competitor changes, and potential opportunities took research across multiple sources. The business needed relevant findings brought together regularly and practical ideas for improvement.

What changed

Scheduled research brings market developments, potential opportunities, and improvement ideas to the team for review. Findings are retained with source links, and subcontractor prospects are organized for outreach and qualification.

What I built

Research and business review workflows with defined sources, report formats, scheduled delivery, and failure checks.

• Daily industry and opportunity research, plus alerts for meaningful competitor changes.
• Weekly researched LinkedIn post ideas and monthly business improvement recommendations.
• A subcontractor research workbook with 246 company/state entries and fields for outreach, qualification, and follow-up.

Prospects still need qualification. Content and recommendations remain subject to human review.

The challenge

A working regenerative farm had about 20 large containers of herbs, spices, and apothecary products still in their original bags. Without an inventory or prices, the owner struggled to find and sell them—and had little time to sort them.

What changed

Customers began buying the products the following week, and the farm continues selling them through Square. The AI-assisted inventory and pricing took about a day; the full effort, including organizing and displaying roughly 200 jars, took close to a week.

What I built

• Photographed every product and used AI to read labels and build a master inventory. Researched supplier and competitor prices by ounce, reviewed the results, and imported the inventory into Square.
• Transferred the products into large mason jars, created clear labels, and helped build and stock shelves in the store.
• Compiled a researched, reviewed customer reference book covering each item’s uses, benefits, and risks.