A real estate agent with no technical background spent 11 months using artificial intelligence to build a hyperlocal educational website — aiming to replace what she calls a fragmented online landscape for buyers and sellers. Irina Norrell, who leads Compass's Irina Norrell Team in the Washington, D.C. metro area, told HousingWire she grew frustrated with clients spending initial meetings covering basic market mechanics. "The need to explain the basics isn't because my clients aren't informed — they're some of the most educated people in the country — but because our industry does a terrible job of explaining itself," she said. After a failed attempt to hire a developer, Norrell began editing the backend herself. When manual efforts began to hit roadblocks, she turned to Claude, Anthropic's AI tool, and now describes her relationship with the technology as a true partnership. "A lot of people look at me weird when I say that working with Claude is a partnership, but it really is," Norrell said. "You cannot just blindly take everything that Claude or any other AI gives you and accept it and just put it on your website. Of course, with the coding and SEO and calculators, I rely completely on Claude, but I do now know a little bit more about all three, so I've managed to progress there." Since launching the site, Norrell said she's been brainstorming ways to stand out from larger real estate portals. "People are not used to having so much information and such detailed information on very local websites," she said. "They can be Googling, say, 'What are the closing costs in Washington, D.C. area?' My website has calculators but also gives you the exact laws and all of those things, I'm still not going to come up on top of searches, just because all the bigger brokerage sites or portals will be on the top." A moment of inspiration: Norrell said she was on the verge of abandoning work on the site after another 12-hour day. She was turning off her computer when she saw Claude's app still open. Until that moment, her use of AI had been limited to writing and researching. But she had kept reading about Claude being very good at coding. "I was thinking, 'Coding, coding,' and an idea came to me," Norrell said. "That section I'd been fighting with all day, where one line of text kept sitting misaligned for reasons I couldn't figure out — what if I showed Claude what I was trying to do and it could code it correctly?" She found an HTML widget, asked Claude to explain how to use it and then requested they code the problematic section together. She copied the code, pasted it and kept working past 1 a.m. "My house was quiet except for my loyal pug snoring up a storm at my feet," Norrell said. "I was exhausted — emotionally, physically — but I knew I wasn't done. I wasn't going back to the developer. I knew I might just pull this off and create the site I envisioned." Looking back after launching the site, Norrell said her biggest mistake was starting too big. "I started with this huge vision, and it just kept getting bigger and bigger," she said. "It doesn't have to be this way. You can add on later. You can start the process and start showing people your value and the value of your team and have a basic website and just add to it. I did it a little bit the other way around, and it was not the easiest way, for sure." Client education post-NAR settlement: Norrell said lawsuit settlements changing how buyer agent commissions are communicated has made transparency more urgent and more welcome. "All of us — clients and agents — were always, not ashamed, but it was always almost like a taboo to talk about your commission," she said. "This is something that is so much money that was basically talked through very fast; either you agree to this or you don't agree to this. I welcome new transparency tremendously. Although sellers and buyers still are kind of confused about how all this works and how the buyer's agent commission now become...
The Big Picture
Norrell's story is a microcosm of a larger shift: AI is democratizing the creation of digital tools in real estate, an industry where fragmented information has long frustrated consumers. According to the National Association of Realtors, 97% of home buyers use the internet in their search, yet local market data often remains scattered or buried behind paywalls. Norrell aims to fill that gap with a site that not only explains market mechanics but also offers calculators and hyperlocal legal specifics.
Her decision to use Claude for coding — despite zero prior experience — underscores how AI is lowering technical barriers. Instead of hiring expensive developers or relying on generic templates, agents can now build custom tools that address specific client pain points. This approach not only cuts costs but also enables a level of personalization that national portals like Zillow or Realtor.com cannot replicate at the neighborhood level.
“"You cannot just blindly take everything that Claude or any other AI gives you and accept it and just put it on your website."”
By the Numbers
- Development timeline: 11 months to build the hyperlocal educational website from scratch.
- Work intensity: 12-hour days, often working past 1 a.m., before the AI breakthrough.
- AI usage: Initially limited to writing and research; later expanded to coding, SEO, and calculator creation.
- Client profile: Serves the Washington, D.C. metro area, where clients are among the most educated in the country.
- Search visibility challenge: Larger brokerage sites and portals consistently outrank hyperlocal sites despite deeper content.
Why It Matters
Norrell's journey is more than a personal triumph; it signals how AI is reshaping real estate intermediation. Traditionally, agents relied on MLS feeds and national portals for visibility, but the ability to create authoritative hyperlocal content gives them a competitive edge. In a market where commission transparency is under scrutiny following the NAR settlement, offering digital tools that demystify costs can be a powerful differentiator.
The winners here are agents who embrace AI to add value, while losers may be those who cling to old methods or depend solely on third-party portals. Large portals, despite their SEO dominance, could see their relevance erode if local agents successfully position themselves as go-to sources for granular market data. However, the challenge remains: without massive marketing budgets, outranking giants like Zillow in search results is an uphill battle.
What This Means For You
For real estate agents, the takeaway is clear: AI is no longer just for tech-savvy teams. Tools like Claude enable any agent to build custom websites, calculators, and educational content without coding. The key is to start small, as Norrell admits, and iterate based on client feedback.
- 1Identify a recurring client question: Pick a common pain point (e.g., closing costs, local regulations) and build a dedicated page or tool that answers it thoroughly.
- 2Treat AI as a partner, not a crutch: Review and customize AI-generated content; local expertise and a personal voice are irreplaceable.
- 3Track and optimize: Monitor traffic and client inquiries generated by your content; refine based on what resonates.
What To Watch Next
The next milestone for Norrell will be whether her site can gain organic traction despite portal dominance. She will also be watching how AI tools evolve — Anthropic and OpenAI release updates frequently that could further improve coding and SEO capabilities. Additionally, regulatory changes post-NAR settlement could increase demand for transparency, benefiting agents who have already invested in digital education.
The Bottom Line
Irina Norrell proves that AI can level the playing field for independent real estate agents, enabling them to build sophisticated digital tools without technical expertise. Her hyperlocal site is a blueprint for how transparency and education can differentiate an agent in a market dominated by big portals. The next frontier: getting those sites to rank on page one — a challenge that will require both creativity and persistence.
