Say you run a 12-rep solar sales org. Call it Meridian Solar. A homeowner fills out a Facebook lead form at 8:47pm on a Tuesday: roof photo, utility company, average bill. Three other solar companies bought that same platform, that same zip code, that same ad slot. Whoever calls first usually gets the appointment, because by the second call the homeowner has already told someone else their roof type and bill range, and repeating the same conversation twice isn't worth it to most people.
Meridian's leads sit in a queue until Wednesday morning, when a setter works through Tuesday's backlog in the order it came in. That's a routing failure, not a staffing one, and it doesn't take a new hire to fix.
What "speed to lead" is actually measuring
The number worth tracking is the gap between lead-created and first-contact-attempted, in your CRM's own timestamps. Not first contact made — attempted. A missed call at 90 seconds beats a connected call at nine minutes, because the missed call still puts your name on the homeowner's phone before the next company's does.
Set a target, not a hope: first outbound attempt inside 90 seconds of the lead landing in the CRM, every hour of every day the ad spend is live. If the ads run until 9pm, the response system has to run until 9pm too. A form that fires at 8:47 and gets called back at 8am the next day didn't miss by an hour. It missed by twelve.
The trigger: connect the ad platform to the CRM, not to a person's inbox
At most solar orgs, the path is the failure, not the setter. Facebook lead ads, Google Lead Form ads, and the website's own contact form each notify someone by email. That email sits in an inbox until someone opens it. Every step in that chain is a place speed dies.
Replace it with a webhook: the ad platform posts the lead directly into the CRM the moment it's submitted, and the CRM fires an automation the same second. No person has to notice, open, or forward anything. From there, an AI agent takes the first pass: it texts the lead within the 90-second window, introduces itself by name and company, and asks one question to start the qualifying conversation. Text first, not a phone call. Most people screen a call from an unknown number at 8:47pm and read a text.
Hi {{first_name}}, this is Ava with Meridian Solar — got your info from
the form you just filled out about solar for your home on {{street}}.
Quick one: is the roof owned by you, and is it a shingle roof or
something else (tile, metal, flat)?
That single message does two things at once: it confirms a real person is behind the lead, and it starts qualifying before a human ever touches the file.
The qualifying script: three facts, disqualify fast
An AI agent qualifying a solar lead doesn't need to sound like a rep working a script. It needs three checkable facts, gathered in whatever order the conversation naturally goes.
Roof ownership and type come first: renters, and roofs too old or too shaded for panels, are a different pipeline from a straightforward reroof-and-solar candidate — sometimes no pipeline at all. Second is the utility provider and a rough monthly bill. A $60/month bill in a low-rate utility territory rarely pencils out; a $280/month bill almost always does, and the agent can route those differently before a rep spends 30 minutes on a call that was never going anywhere. Last is timeline and who's actually deciding — is the person texting back the homeowner, or a spouse who needs to loop someone else in before anything gets booked.
Bad-fit leads shouldn't disappear silently, and they shouldn't consume a rep's calendar slot either. Route them to a lower-touch follow-up sequence (email plus a monthly check-in text) and keep the qualified leads moving straight to booking.
The booking handoff: onto a calendar, not into a queue
The last step is where most speed-to-lead setups quietly fail: the AI qualifies well, then hands off to "a rep will call you," which reintroduces the wait it just eliminated. Instead, the agent should offer real calendar slots for reps who are actually available in that lead's territory and book the appointment directly. Then it texts a confirmation with the rep's name and a photo.
The rep walks into that call already knowing the roof type, the bill range, and the timeline. The qualifying conversation becomes their prep notes, not a transcript to read live.
That handoff is also the point where you decide how much the AI owns versus a person. Some orgs want a human setter reviewing every booked slot before it's final; others let the agent book directly and treat overbooking as a solvable edge case rather than a reason to slow the whole system down. Neither is wrong — the trade-off is speed against a small rate of rework, and it's worth deciding on purpose instead of by default.
What this costs to run
You don't need a call center to run this. A CRM that can receive a webhook, an automation that fires a text within seconds of that webhook landing, and an AI agent connected to both — one that can hold a short qualifying conversation and read a real calendar — cover it.
The build effort is closer to a focused week than a quarter-long project. The case studies on our site show actual build times for comparable AI agent work — not because solar is identical to those industries, but because the mechanics are the same regardless of what's being sold: webhook in, agent qualifies, calendar booked.
The ongoing cost is mostly attention, not headcount: someone has to review a sample of the AI's conversations weekly, catch the questions it's answering badly, and fix the script. Skip that step and the agent drifts — it'll keep sending the 8:47pm text just fine, but the qualifying questions will slowly stop matching what your reps actually need to know before they walk into an appointment.
Being findable when someone asks an AI assistant who to call is a separate problem from what happens after they fill out your form — that's the AI search visibility question covered in the first post in this series. Speed to lead only pays off on demand you're already generating. It doesn't create leads; it keeps you from losing the ones you paid for.
FAQ
Does 90 seconds actually matter, or is 5 minutes close enough?
Treat 90 seconds as the target you design toward, not a hard line where 91 seconds fails. What matters is being first, and in a market where three competitors bought the same lead, "first" is usually decided in single-digit minutes, not tens of minutes. Design for 90 seconds and a slow day still lands you well inside five.
What happens when the AI gets a question it can't answer?
It should escalate immediately rather than guess. A homeowner asking about financing terms, HOA restrictions, or a specific incentive program is asking something worth getting right, and the agent's job is triage and booking, not being the final source of truth on anything with legal or financial weight. Build the escalation path — a text to a live rep, a flagged conversation in the CRM — before you launch, not after the first bad answer.
Do referral and word-of-mouth leads need the same 90-second response?
Less urgently, but don't skip them entirely. A referred lead already trusts you more and is less likely to be calling three other companies at the same time, so a slower response costs less. It still shouldn't sit in a queue until Wednesday. Same-day is a reasonable target for referral leads even without the ad-lead urgency.