For construction companies and contractors, AI lead generation is mostly about keeping the leads you already pay for. Published research shows the odds of reaching and qualifying an enquiry fall sharply within the first hour. The practical system answers every call and form at once, follows up on WhatsApp or SMS, and tracks each estimate in a CRM pipeline, all within US calling and texting rules.
Key takeaways
- •In a Harvard Business Review audit of 2,241 US companies, 37% responded to a web lead within an hour, and 23% never responded at all.
- •The 2007 Lead Response Management study covered more than 15,000 leads. It found the odds of contacting a lead were 100 times lower at 30 minutes than at 5 minutes.
- •For trades and builders, leads usually leak through missed calls on site, unanswered evening forms and estimates nobody chases.
- •AI voice agents and messaging automation close the gap in response time. A CRM pipeline closes the gap in follow-up.
- •In the US, the FCC confirmed in February 2024 that AI-generated voices count as "artificial" under the TCPA. Business texting from 10-digit numbers also requires A2P 10DLC registration.
Why contractor lead generation is a response problem first
A construction company lead generation agency will usually talk about volume: more ads, more directory listings, more shared leads. Volume matters. But contractors who already spend on Google Ads, Local Services Ads, home-service marketplaces and trade directories often have a different problem. They pay for an enquiry and then reach the person hours or days later. By then the homeowner or property manager has spoken to someone else.
For that reason it usually makes sense for HVAC, roofing, remodelling and building firms to fix response and follow-up before adding new lead sources. The aim is to stop paying for leads that go nowhere.
What the research says about speed to lead
Only a few published studies measure lead response directly. They are older than most of the tools now in use, but their findings have not been overturned.
| Study | Sample | Finding |
|---|---|---|
| HBR, "The Short Life of Online Sales Leads" (2011) | Test web leads sent to 2,241 US companies | 37% responded within an hour; 24% took more than 24 hours; 23% never responded |
| Same HBR article, separate dataset | 1.25 million leads at 42 US companies | Firms that contacted a lead within an hour were nearly 7 times as likely to qualify it as firms that waited longer, and more than 60 times as likely as firms that waited 24 hours or more |
| Lead Response Management study (Oldroyd and InsideSales.com, 2007) | Six companies, 15,000+ leads, 100,000+ call attempts over three years | The odds of contact were 100 times lower at 30 minutes than at 5 minutes. The odds of qualification were 21 times lower |
Treat these figures with some care. The 2007 study used data from one vendor's platform across six companies and was not peer-reviewed. Neither study looked only at construction. The direction of the finding is consistent, though: delay costs contact, and contact comes before any estimate.
The same 2007 study also found that call attempts after 20 hours tended to hurt the chance of qualifying a lead. More persistence later does not make up for a slow start.
Why trades and builders feel this more
The way construction and home-services businesses run makes slow responses likely:
- •Owners and estimators are on site. They cannot answer the phone halfway up a ladder or in a client meeting.
- •Homeowners call outside office hours. Evenings and weekends are when people deal with a broken boiler or plan a kitchen.
- •Urgent jobs go to whoever answers. A no-heat call in January or a leaking roof will not wait for a call-back the next morning.
- •Homeowners compare quotes. Many ask several firms at once, so being the first to book a site visit counts.
- •Estimates go cold. A quote sent with no structured follow-up often ends as a lost job nobody recorded.
None of these needs more marketing spend. Each needs a system.
The system: four layers that keep leads alive
Layer 1: AI answers every call and form at once
An AI calling agent answers inbound calls on the first ring, around the clock, in a natural voice. For a contractor it is usually set up to:
- •Greet the caller with the business name and say that it is an automated assistant.
- •Find out the job type (repair, replacement, new build, extension, maintenance).
- •Collect the job address or ZIP/postcode and check it against the service area.
- •Ask about timeline, property type and, where it fits, budget range.
- •Spot emergencies (gas smell, no heat, active leak) and pass them straight to an on-call technician.
- •Book a site visit or call-back into the team's live calendar.
- •Write a summary and push it to the CRM.
Web forms and ad lead forms follow the same logic. Within seconds of a form submission the system sends a text or WhatsApp message to confirm receipt and offer appointment slots. It can also start an outbound call if the person has given the right consent (see the compliance section below).
If you are weighing the cost of this against a human answering service, our breakdown of AI receptionist cost sets out the pricing models.
Layer 2: WhatsApp and SMS follow-up and estimate chasing
Many lost jobs are lost after the site visit, not before it. An estimate goes out by email, the homeowner means to reply, and nothing happens. Automated follow-up keeps the estimate in front of them without an estimator having to remember.
| Day | Message purpose | Channel | Stops when |
|---|---|---|---|
| 0 | Estimate sent, with a summary and a link to accept | Email plus SMS/WhatsApp | Accepted or declined |
| 2 | Check the estimate arrived; offer to answer questions | SMS/WhatsApp | Reply received |
| 5 | Share a relevant project photo, warranty detail or financing option | Reply received | |
| 10 | Ask directly whether they want to go ahead, revise the scope or pause | SMS/WhatsApp | Reply received |
| 21 | Give the estimator a task to call in person | Phone (human) | Outcome logged |
| 45+ | Seasonal check-in or maintenance offer, if the person has opted in | Email/WhatsApp | Opt-out |
On WhatsApp, Meta's rules shape the design. A business can send free-form messages only within a 24-hour customer service window that opens when the customer messages first. Outside that window it must use pre-approved templates, and only to users who have opted in. A WhatsApp automation build handles this by sending templated estimate reminders and switching to conversational replies once the homeowner responds.
Layer 3: a CRM pipeline built for contractors
Generic sales CRMs are built around "deals". Contractors think in jobs, site visits and estimates. A pipeline that matches how the business works is easier for crews and estimators to keep up to date.
| Stage | Entry trigger | Automation at this stage |
|---|---|---|
| New enquiry | Call, form, ad lead or WhatsApp message | AI qualification; service-area check; duplicate merge |
| Qualified | Job type, location and timeline captured | Site visit booking link sent |
| Site visit booked | Appointment confirmed | Reminders 24 hours and 2 hours before; directions sent to estimator |
| Estimate sent | Quote issued from estimating tool | Follow-up sequence starts (Layer 2) |
| Negotiation | Customer asks for changes | Task for estimator; sequence paused |
| Won | Estimate accepted or deposit paid | Job created in scheduling or field-service software |
| Lost | Declined or no response after the final step | Loss reason recorded; long-term nurture if opted in |
| Completed | Job closed out | Review request; referral prompt; maintenance plan offer |
The loss-reason field matters more than it looks. After a few months it shows whether jobs are lost on price, timing, scope or lack of follow-up, and each needs a different fix. Aiwalay builds these pipelines either as a custom CRM or by integrating an existing CRM with the field-service, estimating and accounting tools the firm already uses.
Layer 4: reporting by lead source
Once every enquiry is logged with its source, you can see cost per booked site visit and cost per won job for each channel, not just cost per lead. A marketplace that looks cheap per lead can be expensive per won job if those leads are shared with several competitors. Data analytics dashboards make this visible without anyone exporting spreadsheets.
Lead sources versus what to automate
Each source behaves differently, so the automation should differ too.
| Lead source | Typical behaviour | What to automate | What stays human |
|---|---|---|---|
| Inbound phone calls | High intent; often urgent | AI answering, triage, booking, emergency routing | Complex scoping; upset callers |
| Website contact forms | Medium–high intent; often after hours | Instant text/WhatsApp reply, booking link, call-back if consent was given | Detailed design conversations |
| Google Local Services Ads and paid search | High intent; billed per lead or click | Answering in seconds; logging source and campaign | Disputing invalid leads |
| Home-service marketplaces (shared leads) | The same lead is often sent to several firms | Contact within a minute; booking offered on the first touch | Price conversations |
| Referrals and repeat customers | High trust, higher close rate | Thank-you and referral tracking; maintenance reminders | Relationship calls |
| Builder showroom and show-home visits | Long consideration; finance-dependent | Visit capture, brochure delivery, staged nurture | Sales consultation |
| Commercial bid invitations from general contractors | Formal, deadline-driven | Bid-deadline tracking; document checklist; GC follow-up reminders | Estimating and pricing |
| Past estimates not won | Low cost to re-engage | Seasonal re-engagement for opted-in contacts | Revised quotes |
HVAC, remodelling and home builders compared
"Lead generation for contractors" covers businesses with quite different sales cycles.
| Segment | Sales cycle | What a qualified lead looks like | Automation priority |
|---|---|---|---|
| HVAC and plumbing | Hours to days; seasonal peaks | Location in area, system type, urgency, homeowner or landlord | 24/7 answering and emergency routing; maintenance-plan renewals |
| Roofing and exteriors | Days to weeks; often weather-driven | Damage type, insurance claim status, property age | Fast inspection booking; estimate chasing |
| Remodelling and fit-out | Weeks to months | Scope, budget band, timeline, decision-makers identified | Site-visit booking; design-phase nurture |
| Home builders | Months; finance-dependent | Lot or plan preference, move-in timeline, financing status | Show-home visit capture; long nurture; sales handoff |
| Commercial subcontractors | Bid cycle per project | Project, GC, bid date, scope package | Bid tracking and GC relationship CRM |
For home builders, "qualified leads" usually means buyers with a realistic timeline and a financing path, not just anyone who downloaded a brochure. A system that asks those questions early, politely and consistently gives the sales consultant a shorter, better list.
Compliance: TCPA and A2P 10DLC for US calling and texting
Automated calls and texts in the US are tightly regulated. Getting this wrong carries real legal exposure, so treat the points below as a starting checklist, not legal advice.
AI voice is "artificial voice" under the TCPA
In February 2024 the FCC unanimously adopted a declaratory ruling confirming that AI-generated voices are "artificial" under the Telephone Consumer Protection Act. As Mayer Brown notes, this does not ban AI calls. It means they fall under the existing consent, identification and opt-out rules for artificial or prerecorded voice calls.
In practice, an AI agent answering an inbound call the customer placed is a different situation from an AI agent placing an outbound marketing call. Outbound AI calls need the right prior consent, captured and stored.
Consent rules after 2025
- •One-to-one consent was vacated. On 24 January 2025 the Eleventh Circuit vacated the FCC's one-to-one consent rule in Insurance Marketing Coalition v. FCC. The existing "prior express written consent" requirements still apply.
- •Revocation must be honoured broadly. From 11 April 2025, consumers can revoke consent by any reasonable means, including replies such as "stop", "quit", "end" or "revoke". Requests must be processed within 10 business days.
A2P 10DLC registration for business texting
US carriers require businesses texting from standard 10-digit numbers to register a brand and a campaign that describes the use case, opt-in, opt-out and help flows. Twilio, for example, stated it would block all unregistered US-bound 10DLC messages after 31 August 2023.
| Activity | Rule to check | What to configure |
|---|---|---|
| AI answers inbound calls | Disclosure good practice; state call-recording laws | Opening disclosure; recording notice where required |
| AI places outbound marketing calls | TCPA prior express written consent | Consent language on forms; timestamped consent records |
| Text replies to form leads | TCPA; A2P 10DLC | Registered campaign; opt-in wording; STOP/HELP handling |
| Estimate reminders by SMS | A2P 10DLC campaign use case | Campaign describes reminders; opt-out honoured within 10 business days |
| WhatsApp templates | Meta opt-in and template rules | Opt-in capture; approved templates |
| Re-engaging old leads | Consent scope and revocation history | Suppression list synced across all channels |
The common fix across all of these is a single consent and suppression record in the CRM that every channel checks before sending.
Where to start
A sensible order for most contractors:
- •Measure the leak. Pull 30 days of call logs and form submissions. Count missed calls, after-hours enquiries and estimates with no follow-up.
- •Fix answering first. Put an AI agent on inbound calls and instant replies on forms.
- •Add estimate chasing. Build the follow-up sequence and connect it to your estimating tool.
- •Build the pipeline and reporting. Track every lead from source to won job.
- •Only then add lead sources. Extra spend now goes into a system that converts it.
Use the ROI calculator to estimate what recovering missed enquiries is worth at your average job value. You can also see how this applies across home services.
Further reading: our article on how to calculate lead generation ROI shows how to put numbers on each stage of this pipeline, and the AI calling agents page explains how the answering layer is built.
Sources
- •Oldroyd, McElheran and Elkington, "The Short Life of Online Sales Leads", Harvard Business Review, March 2011
- •Oldroyd and Elkington, Lead Response Management study summary, InsideSales.com, October 2007
- •Cooley, "FCC: AI-Generated Robocalls Illegal Under the TCPA", February 2024
- •Mayer Brown, "FCC Declares Authority and Intent to Regulate AI-Generated Calls Under the TCPA", February 2024
- •Morrison Foerster, "Eleventh Circuit Vacates FCC's TCPA One-to-One Consent Rule", January 2025
- •Manatt, "FCC's Revocation Rule to Take Effect April 2025"
- •Twilio, "A2P 10DLC registration", August 2023
- •Meta for Developers, WhatsApp Business Platform pricing and customer service window