Building AI agents is easy. Building one you can trust with a real business workflow is harder.

I built a Freelance Admin Agent to explore that problem using Strands Agents SDK and Claude Sonnet 4.5 on Amazon Bedrock.

The working MVP takes manually provided client emails and turns them into a structured payment workflow:

→ Classifies invoice and payment emails → Extracts client, invoice, amount, currency, dates and contextual details using structured output → Stores and matches records in SQLite → Uses deterministic logic to identify full/short and on-time/late payments → Uses a second agent to draft a contextual follow-up only when needed → Keeps the final decision with the human through Approve / Edit / Reject

One of the most valuable parts wasn't adding another AI feature. It was finding a real reliability issue during testing: reprocessing the same email could duplicate records and inflate payment totals.

I reproduced it through repeat testing, traced the database behavior, and fixed it.

That reinforced something I'm increasingly applying to my AI engineering work:

A useful AI product isn't just an LLM call. It needs clear boundaries between AI reasoning, structured data, deterministic business logic, persistence, testing, and human control.

The next iteration will focus on stronger production safeguards, including source-grounded financial extraction, precise currency handling, stricter trust boundaries for untrusted input, and eventually live inbox integrations.

I'm interested in building more systems like this for startups and teams with repetitive workflows that can genuinely benefit from AI agents and automation.

Stack: Python · Strands Agents SDK · Amazon Bedrock · Claude Sonnet 4.5 · Pydantic · SQLite · Streamlit

turning client emails into admin tasks is such a real freelancer pain, nice build 🔥

🏢 The Business & The "Greasy Hands" Dilemma: We built DockFlow (Biscayne Marine OS) for Elena "Rico" Rossi, a solo mobile boat engine mechanic operating "Biscayne Mobile Marine" in Miami, Florida. Elena drives a mobile workshop van to marinas and canal slips to repair high-performance boat engines (Yamaha, Mercury). But she faced a brutal daily dilemma: The Greasy Hands Barrier: When working, both of her hands are covered in black engine grease inside noisy boat engine compartments. She physically cannot pick up her phone to answer customer calls. The Traffic & No-Show Trap: Over 35% of boat owners would cancel after Elena had already driven 45 minutes across Miami bridge traffic in her work van. Logistics Chaos: Between texting back and forth for marina security gate PINs, boat parking slip numbers, and checking if ocean tides were too shallow to run boat engines, Elena was losing 14+ hours every week just playing phone tag.

⚡ What We Built in Lovable:

We completely rebuilt the moment an inbound customer inquiry turns into a confirmed, paid booking without Elena ever touching her phone:

Marketing Landing Page (/): An executive business homepage with an interactive 3-slide hero carousel explaining the service, problems eliminated, and dockside repair guarantees.

Customer Fast-Track Booking (/book):

Elena's Mission Control (/owner):

Realtime Auto-Healing Waitlist Waterfall:

Miami Localization & Enterprise Polish:

📈 Measurable ROI: ⏱️ 14 Hours Saved Weekly in manual scheduling and phone tag. 🛡️ Zero No-Shows thanks to upfront card deposits. 🚗 2+ Hours Saved Daily in Miami bridge traffic via neighborhood clustering.

🔧 No Wasted Parts Runs with van inventory pre-flight matching. Live App: https://dockflow-biscayne.lovable.app/

Amazing! Love the backstory and a female boat mechanic with greasy hands. Well done.