December 06 2022

The ugly reality many of us choose to ignore: Not all AI “solutions” are solutions. 

(Don’t worry, we won’t make you select all the boxes that contain a lamp post before reading.) AI, in spite of its best efforts, can have a bad reputation. Teams fear it will take their jobs. Customers hate bad AI and the frustrating experience it provides. In this article, we’ll unpack five reasons why most AI sucks, and show you what to look for when seeking an AI solution that isn’t “the worst.” 

#1. Basic branch logic is the automated circle where good customer experiences go to die

We’ve all been stuck in a neverending chatbot loop. All you want is an answer to a specific question, and all you’re getting is the frustrating unhelpful answers to pre-determined options that don’t align with the question you’re trying to ask. In this AI branch logic world, there is no conversational back and forth, and there are no answers to questions that don’t fit into the pre-determined box.  We’ll call the “looping branch logic circle of customer death” that redirects you to links (with content that you could have Googled), that do not answer your question or help solve your problem. 

For emotionally weighty decisions like where you live, renters need answers to their questions, and if renters’ questions are too detailed…then your solution needs the ability to quickly escalate their question to your 100% human leasing team. 

AI that doesn’t suck: Solutions with customizable handoffs backed by Natural Language Processing (NLP). First things first, handoffs are vital. You need a solution that allows team members to quickly pick up a conversation where it left off and take care of renters’ urgent questions. Solutions built with NLP and powered by machine learning get smarter by answering thousands of questions. The results? A solution that can handle increasingly complex questions with ease, and take many routine follow-up questions off your team’s plate. 

#2. AI feels like a robot

We hinted at this earlier when we bashed branch logic solutions. Where you live is one of the most emotionally weighty and financially significant decisions. Renters don’t want it to feel like a sterile-robotic-emotionless experience. Renters want to * feel * something when they tour your community. They want to know that this is THE place they want to call home. With sucky AI solutions, the conversation can feel robotic, and stilted, which takes the emotional connection out of the experience. 

AI that doesn’t suck:

Can we tell you a not-so-secret story? Many of our clients name their AI solutions human-esque names, AND they also have prospects ask at the leasing office for the AI solution’s name, because they thought that’s who was meeting them to give them their tour. 

Plus, not to brag, but when a community using our AI solution was secret shopped, we earned a 100/100 score hitting every lead nurturing requirement RKW Residential has including immediacy, consistency, personalization, and appointment setting. 

Read the full GlobeSt. Article here. 

#3. AI is biased

One of the biggest challenges in creating an AI solution, particularly in multifamily, is the vast number of questions renters ask, and the nuanced ways that they can ask those questions. (P.S. Branch logic can’t handle this level of nuance and that’s why the branch logic approach above sucks. * steps off soapbox *). 

Instead of building AI logic to account for the many ways messages are phrased, our approach at Funnel relies on machine learning to build a model from a trove of data having worked with many of the largest and leading property management companies in the country. For our AI solution, the data inputs are collections of renters’ messages, which are annotated and indicate their intent and other valuable information. From there, an algorithm optimizes a model that relates the messages to the information they contain, so that when a new message arrives, the intent behind it can be understood by the model. Then, the virtual agent is able to respond successfully to some of the most common inquiries renters make, freeing agents to perform more complex tasks. Because AI runs on data, it can be tempting to conclude that because it’s an algorithm that it is free from bias or human error. That’s simply not true, if there is bias in the data, there is bias in the model and algorithm, leading to bias in the answers. 

AI that doesn’t suck: 

Funnel’s prosocial P.I.N.K. value is literally coded into our products. Read this article to learn more about the steps we took to make our AI unbiased. 

#4. “AI will replace our jobs”

Your teams might fear AI, because they feel that it could replace them. 

We’re of the opinion that this is a very all-or-nothing mindset. AI replaces some of the tasks that onsite leasing team members performed, but AI is built to answer repeatable questions that are asked often, and then when questions become too in-the-weeds AI is trained to escalate the conversation to the human agents. And it’s there to give your teams a hand. 


AI that sucks takes the whole job and does it poorly. AI that doesn’t suck takes the parts of jobs that are mundane and takes care of them without sacrificing quality. Strong AI solutions free your teams to focus on the part of their roles that are high value and can only be handled by a human. Read this Forbes article about how leasing teams can embrace AI.

AI that doesn’t suck: 

We believe that AI amplifies agents’ ability to focus on the parts of their job that are the highest value, and most impactful: assisting renters and residents. We think that AI should save agents time on repeatable administrative tasks. 

As Funnel client, Essex Property Trust shared in a public earning call, they have realized 35% time savings for their team members by leveraging Funnel.

#5 Siloed AI insights

AI isn’t helpful if it doesn’t talk with your CRM, or inform your team of the valuable intel garnered which makes their jobs easier. You need solutions to transparently share information like renter preferences, which gives your onsite leasing team the ability to make their follow-up informed, personalized, and impactful. Renters want to be known, they’re finding their next place to call home. One-size-fits-all impersonal follow-up that doesn’t reflect what you’ve learned about their needs and wants through your AI solution makes them feel like a number, not a neighbor. 

AI that doesn’t suck: 

Your business needs an AI solution that shares the intel gained about a prospect in transparent easy-to-find places. For example, it’s recorded in a single, renter-centric guest card within your CRM. On this guest card, teams can easily access every interaction or conversation that the renter has had with your property management company AND the AI solution parsed their preferences — budget, move-in date, pet information — and recorded it there too. This empowers your teams to communicate with renters in an intentional, personalized manner. For renters, this ensures renters never have to answer repetitive questions or feel like they’re starting at “square one” just because they are talking with a new leasing agent. Win-win. 

Truth: Most AI sucks…but Funnel’s AI isn’t most AI

Bad AI creates poor customer experiences. The quality of the solution that your company chooses directly reflects on the quality of customer experience that you’re able to offer. We built your leasing teams’ new BFF and the solution your renters ask for on tours — from the way we built our products to mitigate bias, to the 100/100 secret shop score we earned from our clients RKW Residential, to the robust natural language processing and machine learning that fuels our solutions — there’s nothing that sucks about our AI.