👋 Hey, Jon here! I have an exciting data set for this new installment of The Data-Driven Trades newsletter. Each month, I will be publishing data from ZyraTalk’s AI CSRs, with benchmarks and comparisons to human CSR performance.
AI is now more than a buzzword, with many real-world, practical applications popping up, but there’s a lot to be desired when it comes to hard data from the performance of those tools.
I admire ZyraTalk for reaching out to us to partner with us in building a dashboard to measure the performance of their AI-driven tools. It’s always a risk to put your money where your mouth is, but as you’ll see below, Zyratalk’s AI CSRs deliver on their asks.
Before we jump in, I have a close family member who is a CSR, and I am aware of some tension and concern about AI replacing this workforce. With any new technology, there’s justified fear, but in working with contractors for half a decade, there is no shortage of work available to improve the customer experience.
The handling of repetitive tasks, or support with call overflow opens a door for higher-level work to improve the customer experience continuously.
The below data is from 8 contractors measuring performance from December 1st, 2024 - January 21st, 2025. This data set will grow over time, but we were excited to get early findings out to you.
6,081 unique leads were tracked in this time frame, with human CSRs handling 70% of the call volume and 30% handled by ZyraTalk AI CSRs.
AI CSRs Drove a 36% Book Rate Across 1,800 Unique Leads
Out of ~1,800 unique leads, ZyraTalk’s AI CSRs converted 36% of those to booked appointments, compared to 39.5% from human CSRs.
The human CSRs handled nearly 4,300 leads and booked 1,825.
Keep in mind, this is a “raw booking rate” - on both the human and Zyratalk’s AI CSR side, no calls are thrown out. There are no “not leads” or “excused/abandoned calls.”
This is a pure metric - total individual leads, regardless of quality, and what percentage of those unique leads turned into booked appointments.
We will have more to come on lead quality as we are developing a product to automate and score on lead quality, as well as CSR scores independent of lead quality, but for now, we are just looking at total leads.
34% of the booked appointments by Zyratalk’s AI CSRs were plumbing services, 28% were HVAC services, and 8% were HVAC installations.
The other 30% was spread out across a few categories, including a small % in water heater installs, other plumbing installs, sprinkler, electrical, and excavation.
It’s clear the Zyratalk’s AI CSRs could handle booking appointments from a diverse array of services.
32% of the booked appointments from human CSRs were from plumbing services, 29% were from HVAC services, 14% were from electrical services, and 6% were from HVAC install. The remaining ~20% was spread out over a few different services.
Of the leads booked by ZyraTalk’s AI CSRs, 17% of them were new customers compared to 40% for human CSRs.
This discrepancy is likely based on the type of leads coming in, e.g. humans just happened to field more new customer leads, but it is another stat to keep an eye on.
Unique Leads Handled By AI CSRs Had A 29.6% Run Rate
About 7% of the unique leads booked by ZyraTalk’s AI CSRs canceled their appointments, compared to about 3.5% from human CSRs.
The difference at this point might be negligible but something we’ll monitor as the data set grows.
The initial key metric we care about with CSRs is the book rate, but we do want to see how those leads fare as they get further down the funnel.
For example, if there was a major drop off (cancellations) between booked job rate and match/run rate, there might be an issue in how those appointments are booked, but both are seeing a normal level of appointment cancellations so far.
16% of jobs run from ZyraTalk’s AI CSR bookings were from new customers, compared to 36% from human CSRs.
It’s unlikely the cancellations were due to the CSR, but it is interesting to note that the cancellation rate amongst new customers was higher from human CSRs than ZyraTalk’s AI CSRs.
Unique Leads Handled By AI CSRs Had A 17% Paying Customer Rate
About 17% of unique leads booked by ZyraTalk’s CSRs turned into paying customers with an average ticket of $3,047.90.
15% of those paying customers were new and spent an average of $4,677.36.
The paying customer rate of unique leads booked by human CSRs was 5% higher, coming in at 22%, with an average ticket of $4,415.05.
Of those paying customers, 34% were new and spent an average of $4,996.53.
Paying customer rates are more reliant on the in-home sales process than they are CSRs, but when doing any analysis, it is important to follow the customer journey all the way through to revenue.
Let’s take a look at the revenue generation comparison between ZyraTalk CSR’s and human CSR’s 👇
Unique Leads Handled By AI CSRs Generated $752,000 of Closed Revenue
Out of the total lead volume in this data sample, ZyraTalk’s AI CSRs handled 30% of that volume, generating $752,000 in closed revenue, or 19% of the total revenue.
Unique leads booked by ZyraTalk’s AI CSRs also generated $1.72 million in revenue opportunity (estimated revenue + sold revenue + closed revenue), which was 17% of the total revenue opportunity generated.
In comparison, the human CSRs, handling 70% of the lead volume, generated 81% of closed revenue for the businesses in this sample.
Although book rates were only slightly different between the two (3.5% higher for human CSRs), the biggest difference were the average tickets, again favoring human CSRs with nearly $1,400 in additional spend per paying customer.
Of the closed revenue generated by ZyraTalk’s AI CSRs, 15% came from new customers, compared to 41% from human CSRs.
Until next time . . .
-Jon