How do HVAC companies with 400 five-star reviews build that number while running 8 crews and no marketing team? They don’t ask harder. They build a system that asks after every job, routes intelligently, and rescues unhappy customers before they reach Google.
TL;DR
- Review velocity (new reviews per week) matters more than total count for local HVAC rankings.
- Most operators lose reviews by asking manually, inconsistently, and without routing unhappy customers away from Google first.
- The Workflow Builder in ServiceAgent runs a post-job sequence automatically: ask, branch on satisfaction, send happy customers to Google, send unhappy customers to a private rescue path.
- No staff required after setup. Every completed job triggers the same sequence.
Why Review Velocity Beats Review Count?
Most HVAC operators think about reviews as a total to build toward. That framing misses how Google’s local algorithm actually works.
Recency is weighted heavily in local pack rankings. A company with 80 reviews that adds 3-5 per week will rank above a company with 400 reviews that hasn’t collected one in two months. Google treats a stale review profile as a signal that the business may no longer be actively serving customers.
The practical implication: you don’t need a massive total. You need a steady flow. That requires a system that asks after every completed job, not just the memorable ones or the ones where the office remembered to send a text.
Review velocity also affects your conversion rate from profile views to calls. A customer choosing between two HVAC companies at 10pm looks at review dates. “Last review: 4 months ago” is a trust problem. “Last review: 3 days ago” is social proof that someone else just used you and was happy.
The Timing Problem Most HVAC Companies Have
Customer satisfaction peaks immediately after a successful service call. The tech fixed the problem, the house is cooling down, and the homeowner is genuinely grateful. That gratitude is your review window.
It closes fast.
By the following morning, the homeowner is back in their routine. The relief they felt has normalized. They meant to leave a review but didn’t get around to it, and now it feels like effort for something that happened yesterday. A same-day request converts at 3-5 times the rate of a next-day request.
The problem isn’t knowledge of this timing. Most HVAC operators know they should ask the same day. The problem is execution at scale.
Your tech finished a job at 7pm on a Friday. Your office manager isn’t sending review requests at 7pm on a Friday. And manually tracking which jobs got a request is work that doesn’t happen consistently on your busiest days. Manual review requests break at the edges: evenings, weekends, peak season, days when the dispatcher is already stretched. Those are also your highest-volume service days.
The timing problem is an automation problem.
What the Standard Review Request Gets Wrong?
Even operators who automate review requests often use a single-path system. Every customer gets the same message with the same Google link. That creates two problems.
The first is wasted potential. A personalized request that mentions the tech’s name and the specific job converts better than a generic “how’d we do?” message. Customers respond when the request feels like it comes from someone who knows what just happened.
The second is more damaging. Sending every customer to Google, including ones who had a bad experience, puts your online reputation at the mercy of whoever had a rough day. An unsatisfied customer who receives a direct Google link is far more likely to leave a 1-star review than one who’s routed to a private channel where the issue gets resolved first.
A good review automation system personalizes the request and routes unhappy customers away from public platforms before they get there.
| Manual Requests | Basic Automation | Intelligent Workflow | |
|---|---|---|---|
| Consistency | Breaks on evenings/weekends | Consistent timing | Consistent timing |
| Personalization | High (when it happens) | Low | High (AI-generated) |
| Routing unhappy customers | Depends on staff judgment | None | Automatic branching |
| Private rescue path | Ad hoc | None | Built in |
| Tracking by tech/job type | Manual spreadsheet | Basic counts | CRM analytics |
| Staff time required | High | Low | Near zero |
## Introducing the Workflow Builder
The Workflow Builder is a visual drag-and-drop canvas inside ServiceAgent where you build automated sequences that run based on events in your CRM and scheduling system. Each workflow starts with a trigger (an event that fires the sequence) and runs through a series of nodes (individual actions the system takes automatically).
For HVAC review collection, one workflow handles everything: request, route, and rescue. It fires when a job is completed, reads the available satisfaction signals, and sends each customer down the right path. Operators who set this up consistently save over 10 hours per week previously spent on manual follow-up and reputation management.
You build it once. It runs on every completed job from that point forward.
The Post-Job Review Workflow: Trigger and Analysis
The workflow starts the moment a job is marked complete in your scheduling system.
The trigger: appointment.completed
What to configure for the trigger?
The appointment.completed trigger fires when a service appointment is marked complete in ServiceAgent’s scheduling module. For HVAC operators, this maps to when your tech closes out a job in the field.
This is the right starting point because it fires at the exact moment your window for capturing satisfaction is open. You don’t want a delay between “job done” and “request sent.” The appointment.completed event is that moment, captured in real time and tied to the specific job, tech, and customer record.
To configure it, open the Workflow Builder, select “New Workflow,” and choose appointment.completed as the trigger type. No additional filter conditions are required unless you want to exclude specific job types (warranty callbacks, for example, where a review request may not be appropriate).
What to check: confirm the trigger fires by completing a test appointment in your system and watching for the workflow to activate in the Workflow Builder activity log.
Node 1: AI Analyze
What it does: The AI Analyze node reads the available context from the completed job record: service notes, any AI voice conversation transcript from the booking or post-job call, and CRM fields on the contact record. It assesses the sentiment of the interaction and produces a satisfaction signal the next node uses to route the customer.
Why it matters: Without this analysis, every customer looks identical to the workflow. The AI Analyze node gives the system the judgment to treat a delighted customer differently from one who mentioned a complaint during the service call. That distinction determines whether the customer gets sent to a public Google page or a private rescue path.
What happens: The node processes available text (job notes, conversation transcript, CRM fields) and outputs a sentiment classification: positive, neutral, or negative. You configure which fields it reads and the sensitivity of the classification during setup.
What to check: Run the workflow on a test job with a complaint note and a separate test job with a positive note. Confirm the AI Analyze output differs between the two. If both jobs return the same classification, check that the correct fields are included in the analysis scope.
Node 2: AI Decision
What it does: The AI Decision node reads the output from AI Analyze and routes the workflow down one of two paths based on the satisfaction classification. Positive and neutral sentiment goes to Path A (the review request). Negative sentiment goes to Path B (the rescue sequence).
Why it matters: This is the branching point. Without it, your automation is a single-path system that treats every customer identically. The AI Decision node is what prevents an unhappy customer from receiving a direct Google review link, and what separates the happy path from the rescue path so each can run its own sequence.
What happens: The node evaluates the sentiment score from Node 1 and branches accordingly. Whether neutral sentiment routes to Path A or Path B is a business decision you set once during configuration. Most HVAC operators send neutral to Path A and reserve Path B for clearly negative signals.
What to check: Test both paths with known positive and negative inputs. Confirm Path A fires for positive sentiment and Path B fires for negative. If routing behaves unexpectedly, adjust the sensitivity threshold in the AI Decision configuration.
Path A: The Happy-Customer Review Sequence
Path A runs when AI Decision routes a satisfied customer to the review request sequence. The goal is to get a Google review while satisfaction is high and the job is fresh in memory.
Node A1: AI Generate + Send SMS
What it does: The AI Generate node drafts a personalized review request SMS using the customer’s name, the technician’s name, and the service type from the job record. The Send SMS node delivers it to the customer’s phone number within minutes of job completion.
Why it matters: “Hey Sarah, glad Mike got your AC sorted out today” converts at a higher rate than “How was your service today?” The personalization signals that this isn’t a generic blast. It feels like the company knows what just happened. AI Generate creates that personalization at scale without anyone writing individual messages.
What happens: AI Generate pulls the customer name, tech name, and service type from the job record and populates a template. A typical output: “Hey [Name], hope the house is already feeling cooler! If [Tech] did a great job today, a quick Google review would mean a lot to the team: [link]. Takes 30 seconds.”
The Send SMS node delivers this within minutes of the trigger firing.
What to check: Run a test job and confirm the merge fields populate correctly. Check that the Google review link goes directly to your review form, not to your profile homepage. The direct link removes one step for the customer and improves completion rates.
Node A2: Wait/Delay
What it does: The Wait/Delay node pauses the workflow for 48 hours before the follow-up reminder fires.
Why it matters: Sending a follow-up the same hour feels like pressure. Waiting a week means the customer has moved on. 48 hours is the window where a customer who opened the first message but didn’t act is still close enough to the experience that a gentle nudge is welcome.
What happens: The node holds the workflow in a paused state for exactly 48 hours from when Path A was entered, then automatically advances to Node A3.
What to check: Confirm the delay timing in the node settings. Check that the workflow doesn’t advance early on the first test run.
Node A3: Send SMS (follow-up)
What it does: The second Send SMS node sends a brief follow-up to customers who received the first message but haven’t left a review yet.
Why it matters: A single request converts somewhere between 10-20% of customers on average. A single follow-up can add another 5-10% without feeling intrusive. The key is restraint: one ask, one follow-up, then stop. No third message, no repeat cycles.
What happens: The node sends a short message: “Hi [Name], just following up on our note from a couple of days ago. If you have a moment, your review really does help the team: [link]. Thanks for choosing us.” No urgency, no incentive language (which violates Google’s review policies).
What to check: Confirm the message matches your brand voice and that the direct Google review link matches the one in Node A1.
Path B: The Rescue Sequence for Unsatisfied Customers
Path B runs when AI Decision identifies a negative sentiment signal. The goal is to resolve the issue privately before it becomes a public review.
Node B1: Human Review
What it does: The Human Review node pauses the workflow and sends an internal alert to the designated team member, typically the owner or office manager. The alert includes the customer’s name, job details, technician name, and the AI Analyze summary of what the complaint signal was.
Why it matters: An automated system can’t resolve a service complaint. What it can do is get the right person the right information immediately. The Human Review node ensures an unhappy customer doesn’t fall through the cracks. The team member sees the alert the same day, with context, before the customer has had time to write a review.
What happens: The workflow pauses and sends an internal alert to the configured recipient with the job summary and complaint signal. The team member reviews the details and approves the next step (the private follow-up SMS) or flags the case for a direct phone call instead.
What to check: Confirm the alert reaches the correct team member on a test run. Check that the job summary included in the alert has enough detail to act on. If notes are sparse, make sure techs know to log the job outcome before closing it.
Node B2: Send SMS (private follow-up)
What it does: Once the Human Review step is approved, this node sends a private message to the customer acknowledging the issue and inviting them to resolve it directly.
Why it matters: A customer who receives a genuine “we want to make it right” message before they’ve posted a review is in a different headspace than one who gets no follow-up. Resolving the issue privately keeps the experience off Google and often turns a dissatisfied customer into a loyal one.
Operators who run this path consistently see measurably better retention from customers who had service issues.
What happens: The node sends a message from the business: “Hi [Name], we heard today’s visit didn’t go the way we’d hoped. We’d like to make it right. Can you call us at [number] or reply here with what happened? We stand behind our work.” No public review request. No generic apology. A direct invitation to resolve.
What to check: Confirm the message goes to the customer, not the internal team. Check that the phone number or reply channel in the message is monitored.
Tracking Review Velocity and Technician Performance
Once the workflow is running, the Workflow Builder analytics give you two numbers worth watching.
Reviews per technician: Which of your techs consistently generates the most reviews? This is usually a combination of service quality and job-closing habits. Techs who mention the review request before leaving (“I’ll send you a quick text. A Google review would really help us”) convert at higher rates. The data tells you who to recognize and what behavior to replicate.
Job-type conversion rate: Which service categories produce the most reviews per completed job? AC repair and installations tend to generate more reviews than maintenance visits, because the emotional relief is higher. Knowing this helps you weigh staffing and scheduling toward job types that build your profile fastest.
Both metrics live in ServiceAgent’s analytics dashboard, tied directly to the CRM data from the workflow. No manual tracking required.
Why ServiceAgent Handles Reviews Differently
ServiceAgent’s 24/7 AI Office Manager doesn’t just book jobs. It captures the post-job moment automatically, reads satisfaction signals from the service record, and routes every customer to the right path without your team touching it.
The marketing and reputation module runs the workflows described here out of the box, connected to your existing CRM and scheduling system. Operators using ServiceAgent for review management see +20% customer retention over time, in part because the Path B rescue sequence turns service failures into loyalty opportunities before they become public complaints.
No dedicated review software subscription. No manual chase. No missed requests on your busiest Saturday. ServiceAgent runs the system while your team runs the jobs.
Frequently Asked Questions
How many Google reviews does an HVAC company need to rank locally?
There’s no fixed number. Competitive metro markets may need 100+ reviews to place in the top 3 local results, while smaller markets can see strong placement with 30-40. More important than the total count is your review velocity (new reviews per week) and your average star rating (anything below 4.2 is a trust problem in most markets). Start by matching the review count and weekly velocity of whoever ranks first in your service area.
Is it against Google’s rules to automate review requests?
Sending automated review requests is permitted under Google’s policies. What’s prohibited is incentivizing reviews (offering discounts or gifts in exchange) and review gating (showing customers a satisfaction screen and only sending happy ones to Google while blocking unhappy ones entirely). The workflow here routes unsatisfied customers to a private rescue path. They’re not blocked from leaving a public review. They simply receive a follow-up first, and that distinction keeps the workflow within Google’s guidelines.
What do you do when an HVAC customer leaves a negative review?
Respond within 24 hours. Don’t defend the work, don’t argue, and don’t copy-paste a template response. Acknowledge the specific issue by name (“We’re sorry your AC was still running warm after our visit”), offer a direct path to resolution (“Please call us at [number] so we can get a tech back out”), and keep the response short. A professional, specific response to a 1-star review is often more persuasive to prospective customers than the negative review itself. It shows you take service seriously. The Human Review path in the workflow is designed to catch these situations before they reach Google, and for the ones that get through anyway, your response speed and specificity matter more than the rating itself.