How I Hired AI: Being AI-First Is a Staffing Decision

Last week, in No team, no platform. No platform, no capital., I told you I run the third-party model: I keep overhead low. I rely on outside firms to get the work done. What I did not tell you is how I planned to make that model scale.

In 2022, before Eternal Companies existed, I decided the firm would scale 2 ways: with online content, and with AI. The content became The Eternal Edge, which you are reading. This letter is about AI.

Here is the belief I want to challenge this week. You believe being AI-first means buying software. A license, a chatbot, a pilot your team abandons by March. I have watched plenty of firms do exactly that and change nothing.

Being AI-first is a staffing decision. And here is the truth of how it runs at my firm: I do not treat AI as another third party. I treat it as internal staff.

The difference matters. A third party is somebody else's team on somebody else's clock. I wait on them, and that wait is the cost of the model I chose. Internal staff answer now, hold the full picture, and carry the work the way an employee would. Last week I wrote that an in-house team buys maximum control and maximum speed at maximum overhead. AI gives my firm the control and the speed without the salaries.

I did not arrive here in one step. It took 3.5 years, and the early part probably looks like your own path so far.

I started with plain ChatGPT in January 2023: questions, drafts, summaries. Then I started keeping each deal in its own ChatGPT project. The documents and the prior work stayed together, so I was not starting over with every new chat. Then came the agentic tools. They do not just answer questions. They execute steps. Today I run AI agents that take entire workflows, start to finish, and come back with finished work.

Through all of it, I kept my outside firms. I still use freelancers. I still pay for professional services: legal, accounting, engineering, brokerage. What changed is one specific category. From day one I was deliberate about not hiring analysts right away. The reports, the summaries, the daily number-pulling every platform runs on went to outside services first. That work has now largely, though not entirely, moved to AI. It came in-house without a hire.

The proof is the dashboard I use to manage my 2 hotels in Georgia.

At 4:40 each morning, both properties' overnight reports arrive on their own. Nobody sent them. At 5:15, the staff I built reads the guest communications, checks the review sites, rebuilds the operating numbers, and publishes everything to a single private page. On Mondays it refreshes the capital plan. At 8:06, that same staff checks that the morning work ran, repairs the most common failure by itself, and only contacts me if something genuinely needs an owner's decision.

By the time I pour coffee, I am looking at one screen that used to be dozens of reports, spreadsheets, and files in my inbox. I do not chase the reports anymore. The report is now a live webpage and updates daily.

Before this existed, that was analyst work. A person pulling numbers, assembling the pack, checking it, sending it, every day. At market rates, paying a person to do that work every day runs into 6 figures a year. Now my overhead does not have to carry that cost.

I have written about AI in this letter 3 times before: 90% of CEOs Say AI Has Done Nothing, The Real ROI of Agentic AI Is Capacity, and Building a platform is your biggest defense against AI. These letters summarize my perspective on AI and how to make the technology work for you. What you just read is what those arguments look like in practice.

You can start the same way in any asset class. The starting point is not a dashboard. It is the report you already rebuild by hand every week: the rent roll, the T12, the sales report, the investor update.

1. Pick one report. Use the report you generate most often. Not the fanciest one. The most repeated one.
2. Automate data retrieval. Set up one folder where the source files land on their own: the nightly report, the bank export, the rent roll. Then point AI at that folder to process what arrives. If someone still has to download files and paste numbers, fix that before anything else.
3. Build one page. Have AI assemble the results into one page you read, instead of 10 files you open. The report is not the win. Never assembling it again is.
4. Audit the data. Do not trust a report because it looks finished. Have a second process check the numbers against the source files, and spot-check them yourself until the system earns your trust. I audit this work the same way I would audit a new analyst's work.
5. Create stress test triggers. Decide what should worry you: a number out of range, a report that did not arrive, a cost that jumped. Have the system alert you the moment one of those trips, and stay silent otherwise.

Then repeat with the next piece of work, one at a time.

The limits are real, and I wrote an entire letter about them: AI works on records, and records are old the moment they are published. The property walk, the relationships, the judgment about what a number actually means, and every final decision stay with me.

I am building this platform through higher interest rates, higher inflation, and a tough capital raising environment. Making a deal pencil, or landing a good client, takes more work than it used to. The staff I built is how I do that work without adding payroll. The more capacity it gives me, the more value I can pass on to my clients, partners, and customers.

This does not mean I never hire again, and this letter is not a case against analysts. I could use one right now. My highest and best use is still relationships and doing deals, and a great analyst would buy me more of both. The analyst is also just one role AI is changing. I see my firm hiring multiple people who are AI-native, because AI still requires humans: hand-holding, orchestration, quality control, and maintenance. Every hire will need that skillset, sized to the role.

When I hire an analyst, it will not be to crank out spreadsheets. That person will help strategize, create, monitor, and quality-check a team of AI agents to scale the business. That is a fork in the analyst career path that did not exist 3 years ago. One branch competes with the machine to produce the same work. The other branch architects and oversees the machines to produce the work.

This is not a technology letter. It is a platform letter. A platform is a team, the systems around it, and the capital story they support. When an investment committee looks at yours, it still asks the same question it asked last week: can this team execute the plan, and can the portfolio afford this team? Work done by staff you built, not staff you hired, is what changes that math. The work gets done, and the payroll your portfolio has to carry stays small. AI is not free. I pay for software, upkeep, and my own time when something breaks. But the math is not close.

Being AI-first is not a software decision. It is a staffing decision. The tools will keep changing. The question will not. Which work needs a hire, which work needs an outside firm, and which work belongs to the staff you build?

The future is clear. AI agents are here to stay, and we all need them on our teams.

I made my call in 2022: the analyst work would not be a salary. Today it is internal staff, built, audited, and on duty before sunrise. The dashboard shows me its work every morning at 5:15.

-Damon

Damon C. Healey, Founder, Eternal Companies I help proven real estate operators build the institutional platform that makes capital come to them.

If you want to pressure-test your platform against an institutional standard, that is what a Platform Edge Session is built for. Book a Platform Edge Session | Get the 2026 IC Stress Test​

P.S. AI has become the backbone of how Eternal Companies runs, and this letter showed you one piece of it. There are 2 more I have not shown you. Those letters are coming, and they land by email first: eternalcos.com/the-eternal-edge​

​
​

​
​

​