Real estate platform provider, myTheo, was exploring strategic growth opportunities and looking for creative ways to increase revenue without requiring additional locations or facilities. After attending a generative AI event hosted by Mission Cloud, they decided to further explore potential machine learning solutions.
With the help of MAP funding through AWS, Mission Cloud was able to develop three new revenue-generating solutions. These solutions allow myTheo to provide its customers a unique offering that sets it apart from competitors and creates a proprietary advantage. The company’s stakeholders have expressed delight with the solutions and are now securing funding and preparing for the next phase of the migration acceleration program.
While exploring strategic growth opportunities, myTheo recognized the challenges associated with geographic expansion. The company was looking for creative opportunities to increase revenue by offering greater value to its customers without requiring additional locations or facilities. After attending a generative AI event hosted by Mission Cloud, myTheo decided to further explore potential machine learning solutions.
At this stage, there was no clear directive beyond finding new ways to drive revenue. myTheo wanted to lean on Mission Cloud’s technical knowledge and experience to flesh out ideas and develop a plan to expand myTheo’s offering. While the customer had a lot of ideas for the future, they didn’t have a roadmap or the internal expertise they needed to execute those ideas.
While discussing opportunities with the customer, Mission Cloud determined myTheo was eligible for MAP funding through AWS. The funding created an excellent opportunity for the customer to learn what was available and create and present business use cases to secure internal and external funding.
Finding ways to create value and increase revenue is a broad challenge that most businesses face. Because of the open-ended nature of the customer’s pre-sales calls, Mission Cloud would need to start with a general approach to find the most effective solution. Max Goff, a Data Science Consultant at Mission Cloud, played a pivotal role in this approach, bringing a wealth of knowledge and experience related to MLOps, MLS and real estate.
One of the possibilities myTheo wanted to explore was how it could have a natural language interface to MLS, such as allowing customers to search or ask questions about listings. The company also wanted to find other ways to create greater value than a standard MLS interface. For example, if a user finds a particular kitchen they like, they can easily search for listings that have kitchens with similar attributes.
A third idea myTheo was considering was how it could create something that allowed agents to personalize newsletters that could be sent to potential buyers and sellers. Relationships are key to success in the real estate industry, and myTheo wanted to offer a way to add more value to a tool real estate agents already rely on for building and maintaining their networks.
The Mission Cloud team leveraged this opportunity to experiment, search for open source solutions and learn what can be applied to an overarching revenue strategy. After considering the potential complexity associated with leveraging a SQL query and genAI platform, Max explored ways to search without relying on structured data.
For myTheo’s first request, Mission Cloud proposed a solution that involved taking all of the VARCHAR, or text data in the SQL tables, extracting it from all of the active listings overnight and combining it into one large string. This string was then vectorized, or turned into numbers, and subjected to cosine similarity searches. The process allowed for a usable natural language query into the data and more cost-effective searching, as it eliminated the need for a language model.
The customer’s second request involved improving search functionality. Mission Cloud identified an open source model that could automatically recognize images and extract text. For example, an image from a kitchen could include green cabinets, which are automatically identified and tagged. This solution was an ideal match for the customer’s needs, working right out of the box. Using open source models, Mission Cloud was able to quickly download and build infrastructure around the model to demonstrate the potential solution.
Finally, Mission Cloud leveraged retrieval-augmented generation and an LLM to address myTheo’s request for personalized newsletter capabilities. The solution would provide myTheo customers with a front-end for prompting the language models to personalize content around the specific characteristics or requirements of a given agent.
Though this project combined multiple elements and capabilities, Mission Cloud worked to develop solutions for each of the three options over a short engagement. In all three cases, Mission Cloud demonstrated the ability to provide value above and beyond the data in the MLS database.
A core focus throughout the project was that myTheo didn’t want to roll out a new feature or tool that was similar to offerings from competitors or other real estate companies. While MLS is often used as a foundation to create map overlays or other common tools, myTheo now has a unique offering that sets it apart and offers a proprietary advantage.
With three well-designed solutions in hand, the company is securing funding and preparing for the next phase of the migration acceleration program, mobilization. Customer stakeholders, from myTheo’s owner and CEO to their tech team, have expressed delight in the revenue-generating solutions Mission Cloud delivered quickly.
AWS Services Used
During the engagement with Mission Cloud, myTheo leveraged multiple AWS and Mission Cloud services to develop revenue-generating solutions, including: