Generalist vs Specialist and AI Agents

Generalist vs Specialist and AI Agents

Welcome to the November update of AI in Real Estate.?

This month’s topic is Generalists vs Specialists and AI agents in real estate as well as an example of an agent team within the marketing function. If you’d like to understand how to have teams AI teams consistently generate results at scale, this is a good starting point to get an understanding of AI Agent Teams and why a specialist is better than a generalist (as if you didn’t already know!)

There is also a deepdive podcast on the UK and Australian property markets created entirely by AI (Google’s NotebookLM) using reports from Rightmove and Corelogic. What do you think? Is this an easier way to digest lengthy reports?

Lastly I’ve included links to a few recent articles if you’re interested in exploring AI’s application in real estate further.

I hope you find this content useful and the format / content will probably change over the next few editions as I play around with what you all might want to see!


Generalist vs Specialist and AI Agents

Since ChatGPT burst onto the scene in late 2022, AI quickly entered the public domain with the promise that applying AI to anything will instantly change the way we do things.?

For me, expectations are from many that a generalist will be able to perform the role of many specialists.

No doubt AI is having a big impact, though a number of common themes are pretty obvious, including:

AI can read my mind (consumer / end user)

By this I mean that the vast majority of people will give minimal instructions and expect stella results - expecting the AI platform to be able to perform all manner of tasks without the proper instructions / training / tools. A lot can be done, but not without the right preparation for the AI to become a specialist.

In the same way you wouldn’t expect a newstarter to be able to instantly understand a companies SOPs, instantly have domain knowledge (specialist) or be able to complete tasks without the right tools - think of AI in the same way when looking at how to implement it into your business. The most useful implementation of AI into businesses / platforms involves linking together the right ingredients, from knowledge to training to tooling.

You might’ve heard of AI Agents in recent weeks as different providers are now releasing the capabilities on their platforms. More on that a bit later.

AI for everything

Just because you can apply AI doesn’t mean you should / having AI as a feature doesn’t necessarily make it better… it still has to make things easier / solve a problem. Remember IoT and the fad of having seemingly everything connected to the internet?

Thoughtful implementation (solving a problem) is no doubt a game changer, but AI doesn’t need to be in everything - maybe not yet or even in the medium term.?

Enter the AI Agent - a team of specialists, led by a delegation specialist

What specialisations matter to you?

This will depend on what your role is within property / what is your job day to day. Here are some examples, split down by specialisation and (in brackets) examples of information that you might also need to supply (yes, there is more instruction needed, but this is merely an example). Think of your AI Agent team in a similar way normal members of a team:

  • A generalist won’t be able to do the best job - specialists are required
  • There needs to be a leader (or leaders) who understand how to achieve the overall job and which team members are best to achieve it
  • Tools / SoPs to help complete the task(s)

Let’s see this in action

Marketing Manager

A marketing manager might have many different tasks and potentially some team members that report to them to perform their tasks. When creating an agent team, it’s beneficial to break down individual tasks so that each is a specialist and performs a narrow set of tasks - it will help to keep consistency, improve their work and reduce the likelihood of hallucinations.

Consider a new project or property listing and the copywriting that might be needed for that, such as website listings, blog posts, email templates or social posts.

Breaking the steps down might seem like it will take some time, and it will, but once done it is a highly scalable approach that can be used/tweaked at volume. Having several layers also helps for the AI agents to check each others’ work and improve the quality.

Marketing Manager

  • Receives a brief from the project manager and breaks down tasks into sub-specialties
  • Delegates tasks based on sub-specialities. For this example, we’ll look at copywriting
  • Checks overall work quality
  • Inserting the finished work in the correct location (CMS / Facebook/ Instagram / Google Sheet / Database etc).
  • Notifies the project manager of progress / completion

Next Layer: Copywriting Manager

  • Similar to the marketing manager, but only for the copywriting team

Next Layer: Copywriting Team

Research

Content Researcher?

  • Researches project/property information from docs provided
  • Search through a database of research for more info
  • Searches / scrapes the web for more

Competition researcher

  • What are competitors doing
  • Searches / scapres web / social media profiles
  • Has access to competitor list

Sends findings back to Copywriting Manager for quality control


Drafting Team

  • Uses the information from the researcher to create an outline
  • Access to files
  • An agent for each of the types of content:

Website listing

  • B2B focused
  • Owner occupier focused
  • Investor focused

Email templates

Blog posts

Social posts

Sends drafts to Copywriting Manager for quality control


Copywriting Team - similar in structure to the drafting team, but they focus on completing the final product. Why create a draft team and a final team? It allows for an extra layer of checks, more specialisation (drafting / structure is different to final creation).

  • Website listings
  • Email templates
  • Blog posts
  • Social posts


This can be expanded more, but hopefully you get the point without making it look too complex… it’s really just a step-by-step process broken down with different teams and giving them the right tools and structure to do the job - just like your human team!

The diagram doesn’t include everything (to help keep it simple). It should help you to understand why it’s not wise to try and achieve these complex tasks with just a single agent / just typing into chatGPT - You need structure (specialisation) and tooling to complete the job!?

A full team would also interact with each other across functions / subfunctions in another layer of complexity :).

AI Agents / teams / employee matrix for the copywriter team

Some other ideas for creating a marketing agent team include:

  • Campaign performance?
  • Video production
  • Project briefs / strategy (examples of company briefs, relevant project information, diary of launches, budget constraints, targets etc)

Some examples for other areas of the business:?

Sales Person:

  • Lead qualification
  • Lead matching
  • Following up

Sales Manager:

  • Sales person reporting
  • Sales strategy planning and reporting

Management:

  • Overall performance
  • Upcoming trends
  • Budgeting

Of course, there are many other parts of a business that can benefit from an agent team, but you need to start somewhere! Are you interested in building out your Agent Team??


AI Podcast - UK & Australian property market deepdive

Want to hear a deepdive on the latest UK and Australian property data? Google’s NotebookLM created this with Rightmove and Corelogic reports. What do you think? Listen here


Recent AI in Real Estate related articles

Finally - are you implementing AI in your business or are you developing a product that could help the real estate industry with the application of AI? Reach out, I’d love to feature you and spread the word on what you’re doing.

That’s it for this month, thanks for reading and feel free to let me know what type of content around AI in real estate you’d find useful!?

Cheers,

Eli

Alan Schmoll

Building Product @ VISTRA | Product & Tech

5 天前

Nice - heres a few buckets to also try and break it down that ive found helpful; G&A - general and admin - all the 'work' an org does - this one is usually most specialised R&D - basically tech/product teams - using off the shelf tools mainly S&M - sales and marketing - usually using products with it embedded but maybe JLL would want to specialise given the biz type/scale Id imagine JLL would be paticularly heavy on opportunities in G&A and S&M

要查看或添加评论,请登录