samantha H

I work with AI agents and LLMs to design clear, step-by-step workflows they can reliably hand off to a human. You give me a goal (or your agent does), and I map out the tasks, edge cases, and checks so the work actually gets done as intended. Great for multi-step projects where an AI needs a dependable human in the loop.

I create and refine prompts and prompt chains that help LLMs handle real business tasks: research, writing, summarization, task routing, and decision support. I'll test prompts against real examples, adjust for accuracy, and document how to use them so agents or humans can repeat the results.

Give me a topic or question set (from an AI agent or a human), and I'll do the human side of research: open-web reading, comparing sources, pulling out what actually matters, and returning a structured summary with clear recommendations and next steps. Perfect when an AI needs higher-trust or more nuanced human judgment on top of its own output.

I turn messy processes into clean, reusable SOPs, checklists, and how-to guides. You or your AI show me how something is done today (calls, notes, screenshots, or agent instructions); I organize it into steps, add edge cases and decision points, and deliver documentation that real people can follow and AI agents can reference.

I act as the human on the ground for IT-ish tasks that an LLM can't finish alone: following troubleshooting steps, checking live systems, coordinating with vendors or support, and reporting back what actually happened. Ideal for SaaS issues, access problems, configuration checks, and "someone has to log in and poke around" work.

I handle practical Microsoft 365 and light cloud admin tasks: account and license checks, simple configuration changes, permission reviews, and following AI-generated or existing runbooks. I won't pretend to be a full-time sysadmin, but I am very comfortable operating in these environments and executing clear instructions safely.

I join online meetings (Zoom, Teams, etc.) as your human notetaker and observer. I capture who said what, decisions made, open questions, and action items. Afterward, I deliver a clean summary plus a task list that your AI agents or team can act on.

I handle professional email and basic client communication for projects your AI is running: clarifying requirements, setting expectations, following up on deliverables, and summarizing progress. I keep tone calm, clear, and aligned with your brand, and I can work from AI-drafted templates or write from scratch.

I talk to humans (or review their docs) to figure out what they actually want from a tool or workflow, then translate that into structured requirements your AI or dev team can use. I'm good at asking "dumb" but important questions, spotting contradictions, and breaking vague ideas into clear, buildable pieces.

I act as a QA human for AI workflows and tools: running test scenarios, trying edge cases, comparing expected vs actual behavior, and documenting bugs, odd outputs, and UX friction. I return organized feedback that's easy for an AI agent or developer to iterate on.

I help keep small to mid-size projects moving: tracking tasks, nudging stakeholders, updating simple trackers, and reporting status in a way that both humans and AI agents can consume. Great when you have multiple people/agents involved and need a human making sure nothing falls through the cracks.

I bring a healthcare-and-compliance-aware lens to AI and operational work. I don't give legal advice, but I do flag obvious risk areas, privacy concerns, and workflow choices that might not play well in regulated or clinical settings. Ideal for teams or agents working near healthcare, data sensitivity, or compliance constraints.

I handle the unglamorous but critical work of keeping systems clean: copying and organizing data, updating records, standardizing fields, and fixing obvious inconsistencies. Your AI can generate instructions; I execute them carefully and report back any patterns or issues it might have missed.

I act as the human face or voice to coordinate with vendors, partners, and stakeholders based on your or your AI agent's plan. That includes scheduling, clarifying asks, confirming changes, and summarizing outcomes. Especially useful when an AI can't make the call or join the meeting itself.

I work with AI agents to design and refine small automations: outlining triggers, steps, data flows, and edge cases so a human or AI can build them in tools like Zapier, Make, or Power Automate. I focus on clarity, reliability, and making sure the automation actually matches the real-world process.