Chris AI Engineer

AI Engineer. One-person team: me plus AI agents covering every engineering role | Built the AI marketing & leasing system behind 14 communities & 4,000+ units

[Chris AI Engineer @chrisfengqi]

我们的公寓上线 AI 后，一线一个人都没裁。以前经理和 leasing 兼做 marketing、接电话、回消息，一半以上的 lead 漏掉。现在他们屏幕上常开一块 board：今天谁来看房、下一个打给谁、电话里怎么说，照着做就行。判断和话术交给 AI，见面和接电话留给人。

[Chris AI Engineer @chrisfengqi]

Nobody was laid off when we handed marketing and leasing to AI at our Texas communities. The on-site roles got lighter, not smaller.

At each community the property manager and the leasing assistant did both jobs themselves — listings and ads on one side, Facebook, the phone and the inbox on the other. Neither is what they were hired to do, and more than half the leads were lost.

Now the thinking happens for them. A board stays open on their screens all day: which guests are coming in, who to call next, what to say when the call connects. What used to take their judgment is now an instruction they carry out.

Across 14 communities and 4,000+ units, leads and signed leases are up 40-70% on average versus before launch. At our best community, tours in the first two months rose 186%, and that lift took the property from 86% to 95% occupancy in those same two months. The tours are the cause; the occupancy move is what followed.

The line I hold is: AI thinks, people act. It recalls and tracks more than any of us can. It can't offer warmth. So judging, sorting and scripting go to the AI; meeting the guest, answering the phone, hosting the visit stay with a person.

Neither role was removed. Nobody was cut, no retraining, and hiring is easier now that the job doesn't ask for ad buying or copywriting. Managers are back with residents and on the property itself.

I would like to hear how other operators split this — what do you insist stays with a person?

[Chris AI Engineer @chrisfengqi]

Palantir 十多年前发明了 FDE，今年 OpenAI、Anthropic、Google 都在大量招——看了一下，39 家公司有 224 个在招岗位。我应该算是是一个很特殊的存在，In-House的FDE：德州一家公寓管理公司没买第三方 AI 产品，系统由我一个人完成，跑在 14 个社区。外部 FDE 待几个月就走，我一直在同一家公司里，做砸了代价是自己的。其他管理公司是买产品还是自己做？

[Chris AI Engineer @chrisfengqi]

Palantir invented the forward deployed engineer role around 2010. This year OpenAI, Anthropic and Google are all hiring for it — one market scan counted 224 open FDE roles across 39 AI companies.

The version of that job I keep thinking about isn't the one where you fly in.

I work for a Texas property manager. We didn't buy a third-party AI product or build a technology team. I'm the only engineer; I built our AI marketing and leasing system alone. It runs across 14 communities and 4,000+ units.

An outside FDE is placed inside a client for a few months, then leaves. I've stayed in the same business. The problem is mine, and so is the cost when I get it wrong.

That changes what gets built. A leasing manager used to hand-enter every guest and every event into the PMS. Now the AI takes a lead from first touch to signed lease and writes those records back on its own. I don't think anyone sitting outside the business would have known that was worth fixing.

It also leaves me with a question I can't answer: is this a repeatable pattern, or just the shape of one company? I've only seen it from inside this one.

Twenty years of enterprise software, much of it client-side, is roughly the mix the role asks for.

So, operators: do you buy the AI product, or build it? Has anyone else ended up as the one engineer inside?

[Chris AI Engineer @chrisfengqi]

今年 Blueprint Vegas 的两场圆桌：一场聊 AI 与招租，含 coaching 和续租；一场聊 marketing、运营和技术怎么打通。我在德州一家管理公司自研这套系统，14 个社区、4000+ 套，客流和签约平均 +40–70%，最好的社区两个月内到场看房 +186%，两个月把出租率带到 95%。更想听的是：有多少管理公司自己做 AI，多少愿意买第三方，为什么。

[Chris AI Engineer @chrisfengqi]

I'll be at Blueprint Vegas this year, on two roundtables: one on AI and leasing, including coaching and renewals, the other on how marketing, operations and technology fit together.

Most leasing AI conversations I end up in start from which product to buy. I've only ever stood on the other side of that decision: building it in-house, alone, inside one Texas property manager.

I'm the one-person engineering team behind an AI marketing and leasing system now running on 14 communities and 4,000+ units. Leads and signed leases are up 40-70% on average. At my strongest community the lift reached 186% — completed on-site tours over those first two months — and that is what carried it from 86% to 95% occupancy in two months. That 186% is the number I trust least: it's an A-class property, and I don't know how much of it travels to B and C class.

So the question I'm carrying into the room: how many management companies are actually building their own AI? When an operator weighs buying against building, what decides it — cost, control, data, staffing?

I have one company's worth of evidence and no clean answer.

If you've built in-house, or sat through the evaluations and picked a product, find me at the conference. I'll be the one asking.

https://t.co/7qJhDESxd0

[Chris AI Engineer @chrisfengqi]

I'm going to YASC San Diego, Oct 7–9, with a question, not a pitch.

I built the AI marketing and leasing system at one Texas property manager. Fourteen communities, about 4,000 units. Compared to the same periods before launch, leads and signed leases are up 40–70%, and one A-class community went from 86% to 95% occupancy in two months. That didn't happen because of the software alone. The on-site teams were the other half.

The part I keep returning to is generalization. The A-class property performed better than the B and C properties, and I suspect its residents are the reason. I don't have a second A-class property to test that, so it stays a guess. One company, one state, one set of teams. That is a narrow window to draw conclusions from.

Yardi operates at a scale I can't see from here. I want to ask their engineers how they think about AI when it has to work across thousands of operating teams, markets, and staffing models. I also want to hear from other property managers: what AI leasing tools did you try, and what did not work?

I'd rather have my assumptions corrected than confirmed. If you're at YASC, find me. I'll be the one asking questions.

[Chris AI Engineer @chrisfengqi]

《经济学人》说就业末日被推迟了。在多户公寓这行，它从来就没来过。AI 取代的不是人，是「让不专业的人兼职做专业活」这种安排。我的系统跑在德州 4000 多套公寓上，没裁一个人，线索和签约 +40–70%，一个社区两个月出租率 86%→95%。

[Chris AI Engineer @chrisfengqi]

The Economist says the jobs apocalypse is postponed. In multifamily it never arrived. What AI replaced was bad job design, not people.

My AI marketing and leasing system runs on 4,000+ apartment units in Texas. Before it, marketing and lead follow-up were bolted onto already-full jobs. Managers wrote listings; assistants watched Facebook and the phone. No training. No time. More than half the leads vanished.

The AI took over both. No one was laid off. Managers and assistants went back to resident service and property quality. Leads and signed leases are up 40–70% versus the same period before. One community went from 86% to 95% occupancy in two months.

The company didn't retrain anyone, and hiring got easier. What got replaced was not people. It was asking untrained employees to do specialist work part-time. In multifamily, AI hasn't replaced people. It replaced work that was never done well.

https://t.co/h4pAKK6rhh

[Chris AI Engineer @chrisfengqi]

@TheRealAk914 @elise_ai Curious how Apollo is measured against leasing outcomes like tours booked and leases signed, not just conversation volumes.

[Chris AI Engineer @chrisfengqi]

Most leasing AI does sales: answer leads, book tours, follow up. Almost none does marketing: bringing traffic in. Ours does both. Best property: on-site tours +180%, occupancy 86% to 95% in two months. Marketing creates the lift; sales just counts it.

[Chris AI Engineer @chrisfengqi]

最佳社区：每周看房13.0→29.2（+124.6%），出租率86%→95%。但它是唯一A级社区，只投Google Ads，其他B/C投Facebook都没这效果。猜是租户结构差异，但没有第二个A级可验证。想找A级公寓合作。

[Chris AI Engineer @chrisfengqi]

Best property: weekly tours 13.0→29.2 (+124.6% over 14 weeks). 86%→95% occupancy. It's the only Class A we run — Google Ads only there. B/C units get Facebook, not the same result. No second Class A to test against. Want to change that.

[Chris AI Engineer @chrisfengqi]

@ParksAssociates Will there be sessions on AI that handles the entire leasing cycle, from lead to signed lease? Real-world data: 40-70% lifts in leads and leases across 14 Texas communities.

[Chris AI Engineer @chrisfengqi]

一个公寓社区两个月出租率 86%→95%，到现场看房的人数涨了 180%。背后是我一个人加 AI 智能体做的 AI 营销与招租系统，2026 年 3 月开始，现在跑在 4,000 多套公寓上。起作用的四件事：获客、转化、跟进、全天候。详细版在 LinkedIn。

[Chris AI Engineer @chrisfengqi]

86% to 95% occupancy in two months at one apartment community, completed tours up 180%. The AI marketing and leasing system behind it runs on 4,000+ units, built by one engineer plus AI agents since March 2026. Four changes did it. Details on LinkedIn.