6 hours to 1.
10-15% margin recovery.
Zero to 42% meeting booking.
A live dashboard in 10 days.

The problem

Running 7 active projects from a 10-tab Excel spreadsheet. Finding out about cost overruns 30 days after the fact. Losing 10-15% of project margins to manual processes the team had accepted as normal. The CEO couldn't answer "are we making or losing money?" without asking three people and waiting for a spreadsheet update.

What I built

Capataz, an AI system that searches across 10+ years of contracts, tracks project costs against actuals in real time, automates certification for compliance, manages workers across sites, and gives the executive team a dashboard showing every critical metric without spreadsheets.

The result

• ▸CEO sees every project's cost and schedule status in real time
• ▸Monday status meetings eliminated, the dashboard replaced them
• ▸Field teams report faster, finance catches overruns 3 weeks earlier
• ▸Document lookups went from 60 minutes to 30 seconds

The problem

12 years of published research, 104 documents, and 77 podcast episodes. None of it visible to AI search. Prospects asking ChatGPT or Perplexity about wealth management never saw the firm. Branded searches worked. Everything else was a 71x gap. Fixing that opened conversations with every partner, and each one independently described the same six operational problems: proposals built from scratch, institutional knowledge trapped in email, meeting transcripts that disappeared after every call, and over $100K in AI tools deployed but never configured.

What I built

Voice DNA extraction from 12 years of firm content. Authority framework across 10 discovery questions prospects ask AI before calling. 60 structured authority pages written in the firm's calibrated voice. AI visibility gap closed from 71x to 2.1x within 90 days. Engagement expanded from a 6-week content project into a proposed 9-month Fractional Head of AI retainer to build proposal automation, a searchable knowledge base, meeting intelligence, and a daily briefing system.

The result

• ▸AI visibility gap: 71x reduced to 2.1x
• ▸AI search presence: from 2 of 10 industry queries to 8 of 10
• ▸Proposal time (projected Phase IV): 4 to 8 hours to minutes
• ▸Knowledge retrieval (projected Phase IV): unsearchable to seconds
• ▸Engagement type: Fractional Head of AI

The problem

Website scoring 8 out of 100 on performance. Leads from Google and Meta ads landed on a page that barely loaded, then went to a shared inbox where attorneys manually qualified every inquiry over WhatsApp. No CRM, no tracking, no way to know which leads converted or where they came from. Most never booked a meeting.

What I built

Victoria, a WhatsApp AI agent that qualifies leads automatically, books meetings into Calendly, and syncs everything to a new CRM with source attribution. Three service-specific landing pages replaced the broken site. Attorneys stopped manual WhatsApp within the first week.

The result

• ▸Meeting booking rate: 0% to 42% of inbound WhatsApp conversations
• ▸Lead response time: hours/next-day to seconds
• ▸Website performance: 8/100 to fully rebuilt with 3 landing pages
• ▸First client for Work-Smart.ai, came from a referral, led to everything else

The problem

Six custom ChatGPT bots, none connected to each other. Every client call meant hours of manual follow-up: summarizing notes, creating tasks, drafting emails. A 4-hour workshop generated 6 to 9 hours of documentation work. The founder and her VA were repeating context to AI tools that forgot everything between conversations.

What I built

Read everything she'd ever written, proposals, strategic plans, emails, workshop decks, and extracted a 280-line communication profile. Structured her entire process into a shared workspace with 15 reference documents and 7 pre-built AI skills (meeting debriefs, email composition, grant blueprints, newsletters, discovery call processing, proposal generation, and workshop creation). Her VA now runs the system independently.

The result

• ▸Grant blueprints from 6-8 hours to 2-3 hours
• ▸Meeting debriefs from hours of documentation to minutes
• ▸VA productivity increased, same person, same hours, 3-4x more output
• ▸First standalone training/workshop client for Work-Smart.ai

The problem

Three divisions (Residential, Office, Revalue) with no unified digital presence. No client portal. No admin dashboard. No CRM connecting leads, projects, and analytics.

What I built

Three optimized websites (one per division), a client portal where projects track milestones and get automatic progress summaries, an admin dashboard that connects leads, projects, and revenue in one view. Included training on how to use AI tools for future operations.

The result

• ▸Unified digital presence across three divisions
• ▸Client portal replaced manual update emails
• ▸Admin dashboard provides single-view operations
• ▸Team trained on AI tools for ongoing use

The problem

Website down half the time. A field sales team across 4 countries with no digital selling tools. Reps carried physical catalogs. No career portal despite constant hiring.

What I built

An integrated digital platform where sales, catalog, customer operations, and lead capture run from one place across all four markets.

The result

• ▸Digital credibility restored across all four markets
• ▸Field sales team selling from one unified system
• ▸Hiring pipeline consolidated into a single workflow
• ▸Inbound leads captured automatically into the same platform

The problem

Running a membership business on separate tools. Website on Squarespace. Membership on Memberful. CRM scattered across email. No content hub. Booking through Calendly with manual CRM entry.

What I built

A unified platform: website, membership billing, content hub, booking system, and CRM all talking to each other. Leads get qualified automatically, booked, and tracked, no manual re-entry.

The result

• ▸Everything managed from one place
• ▸No more manual data entry between systems
• ▸Business owner reclaimed operational hours

The problem

Tens of thousands of SKUs across multiple US distribution territories, all analytics running through Excel pivot tables connected to a legacy on-premise ERP. Getting a sales breakdown by territory, customer, or SKU meant hours of manual filtering. Leadership asked one question and got three different answers.

What I built

A real-time sales dashboard pulling directly from the legacy on-premise ERP. Leadership sees sales by territory, customer, product, and salesperson, in units. Daily automatic updates. Still exportable to Excel for teams who prefer it.

The result

• ▸Report generation from hours of filtering to seconds (page load)
• ▸One shared dashboard replaced individual Excel files across the team
• ▸Number consistency, one source of truth instead of three different answers
• ▸10 business days from kickoff to live dashboard
• ▸Foundation set for AI layer: anomaly detection, purchasing suggestions, inventory alerts