Episode 035 HR, AI and the Proactive Solopreneur
The biggest AI mistake in a small law firm may have nothing to do with legal work. HR strategist Natalie Dalton explains why solo lawyers with outsourced paralegals, offshore assistants, and AI-written handbooks still need a people plan, and why you should diagnose your practice before you digitize it. Plug AI into a broken process and all you get is a broken process that runs faster.
Listen/Watch On
Show Notes
Every legal tech conference wants to talk about AI drafting and contract review. Almost nobody talks about the business behind the practice: the virtual assistant in the Philippines, the outsourced paralegal, the 1099 you found through a Facebook ad. Natalie Dalton, founder of Out of the Box HR, says that's the blind spot. Lawyers keep buying software before they understand how their own firm actually works. Her rule is simple: diagnose before you digitize. So what happens when you plug AI into a process that was broken to begin with?
In this episode:
- Why HR is really organizational design, not just payroll, benefits, and handbooks
- Whether a solo lawyer with an outsourced paralegal and an offshore assistant needs HR at all
- Employer of record arrangements and the compliance risk of hiring overseas workers directly
- Starting with strategic objectives: defining roles, decision rights, and accountability before you hire
- Using ChatGPT and Claude as a planning partner, and why an AI-drafted employee handbook only gets you about 80% of the way there
- AI resume screening: setting must-haves and deal breakers, and keeping human eyes on the shortlist
- Natalie's time at Document Crunch, an early AI platform that put construction contract knowledge in the hands of field teams
- Natalie's AI Decision Tree and Law Firm AI Playbook, built around "diagnose before you digitize"
- Reactive vs. proactive firm owners, and the hidden time bomb of a disengaged employee
- Using AI to structure one-on-ones and turn meeting notes into follow-up action items
We also discuss:
- Why an AI that always tells you you're right is a risky HR advisor, and how Heather's AI assistant Bosley pushes back
- AI memory failures, even with instructions stored in GitHub
- Adding an HR consultant to the classic startup team of insurance agent, lawyer, and accountant
- How Heather tested AI resume screening against a human assistant before trusting it
- The 2026 midterm ballot measures on AI data centers, and whether slowing AI development would be such a bad thing
- AI as "durable infrastructure," and why "AI app" will soon sound as dated as "internet app"
- Voters in 13 states weighed 50 local ballot measures on AI data centers in 2026, and nearly every decided result went against development. Is slowing AI down such a bad thing? Ballotpedia — Data center-related ballot measures, 2026
Download: Natalie Dalton's AI Decision Tree and Law Firm AI Playbook
A practical framework for deciding where AI belongs in your firm before you spend money on software.
- Starts with the business, not the tool
- Identifies where time and expertise are being wasted
- Separates technology problems from management, training, and workflow problems
- Helps you avoid automating a process that doesn't work
Key Takeaway
Natalie's point is blunt: install technology on top of an ineffective process and you just make that process ineffective faster. Before you buy another subscription, figure out whether your real problem is the tool, an unclear role, a broken workflow, or a person nobody is managing.
Mentioned in This Episode
- Natalie Dalton, founder of Out of the Box HR
- Natalie's AI Decision Tree and Law Firm AI Playbook
- Document Crunch
- ChatGPT
- Claude
- Anthropic and OpenAI
- Propel (outsourced paralegal services)
- Indeed
- Bosley (Heather's AI assistant)
- Hugging Face (discussed in [Episode 033])
- 2026 midterm ballot measures on AI data centers
- Jen Grondahl Lee (returning guest)
Highlight Reels on YouTube and Instagram
