How Concrete Crews Schedule Pours Around Weather Without the 5:30am Phone Chaos

The 5:30am weather scramble that runs your whole week

It is 5:30 on a Tuesday and the operations lead for a four-crew concrete outfit is standing in the dark driveway staring at his phone. The overnight forecast flipped: rain now sits over the county from 10am to 2pm, right on top of the 40-yard foundation pour scheduled at the new build. So the calls start. The ready-mix plant, to push the load and hope there is a truck free Thursday. The pump company, because cancelling the boom pump inside 24 hours means eating the minimum. The finishing crew foreman, already hooking up the trailer. The general contractor, who wanted that footing yesterday.

By 7am the day has been rebuilt three times in his head and nowhere else. The office whiteboard still shows Tuesday's dead plan. Jobber has the jobs on a calendar, but Jobber does not know a foundation pour is weather-locked and a garage slab is not, so it just sits there showing the wrong day. The estimator, who should be out measuring a $28,000 stamped patio, is instead on the phone chasing trucks. That patio quote from last Thursday is still sitting in the customer's inbox with no follow-up, and by Friday the homeowner will have booked whoever called back.

None of this is a calendar-app problem. It is a "your business runs on weather and shared equipment and nobody built software that understands either" problem. That gap is exactly where a focused custom build plus a few sharp AI automations pays for itself.

Rain in the forecast, and no system to reshuffle the day

The pain is not the rain. It is that every forecast change forces a manual rebuild of the whole day, from memory, under pressure, by your most expensive people. A pushed pour drags three vendors with it: plant, pump, and crew. Get one wrong and finishers stand around at $45 an hour waiting on concrete that is not coming, or you pour a slab into a puddle you tear out next week. ServiceTitan and Jobber cannot fix this: they are calendars with invoicing bolted on. They do not ingest an hourly forecast, they do not know a footing needs to stay above 40 degrees through its cure window, and they have no concept of which jobs are weather-safe fallbacks.

A custom scheduling engine treats weather as a first-class input. Every job carries its own rules: minimum temperature, rain sensitivity, cure window, whether it is an interior pour that runs in any weather. The engine pulls the hourly National Weather Service forecast for each site's zip code, and when a pour crosses a threshold it flags the at-risk jobs the night before, not at 5:30am. Then it proposes the reshuffle: foundation to Thursday's open plant slot, garage slab pulled into Tuesday, with the exact texts to plant, pump company, and customer drafted so the dispatcher approves instead of dials. The day gets rebuilt once, in software, before anyone is standing in a driveway.

The 9pm call that goes to voicemail and books your competitor

A homeowner decides at 9pm that this is the summer the cracked driveway finally goes. They call the first concrete company on Google, it rings out, they call the next one. That call was worth $12,000 and you never knew it happened. Across the crews we have built for, missed and after-hours calls are the single biggest quiet leak in the business. An answering service does not fix it, because a generic call center cannot ask whether this is a tear-out, whether the truck can reach the back yard, or whether they want stamped or broom finish. It books a hollow "quote," or just takes a message you read the next morning.

An AI phone agent trained on your trade answers every call, day or night, and qualifies the way your best estimator would: slab or footing, rough square footage, tear-out of existing concrete, truck access, decorative or standard. It books the site visit on the estimator's real calendar, texts a confirmation, and writes the lead into ServiceTitan with the notes attached. The outcome the owner cares about: the phone is answered at 9pm and there is a job on the calendar by morning.

The $28,000 patio quote that sat three days and lost the job

The estimator writes a strong quote for a stamped patio, sends it, then the week catches fire and nobody follows up because everyone is reshuffling pours around the weather. Three days later the customer who was ready to sign has gone cold or hired whoever called back. The quote was never the problem; the silence after it was. Your CRM (Customer Relationship Management) stores the estimate but will never chase it: Jobber and Housecall Pro email a quote, mark it "sent," and there the story ends.

An AI follow-up flow watches every open estimate. Two hours after a quote goes out it sends a short, specific text, not "just checking in" but a note that names the stamped patio, answers the real questions about cure time and financing, and offers two windows to get on the schedule. A genuine reply hands off to the estimator with full context. Silence gets a nudge on day two and day four, then the lead is flagged cold so nobody wastes another call. The close stops depending on whether a stressed human remembered to dial.

Two crews, one pump truck, and a double-booked Tuesday

You run two finishing crews and own exactly one boom pump. Working off a whiteboard and a group text, the dispatcher books both crews on pours that need that pump at the same hour, or sends a crew across the county with no travel buffer so they sit idle burning labor while the concrete truck waits. Every idle hour is real money, and shared equipment is where the schedule quietly falls apart. Generic field-service scheduling does not model a single-instance resource: ServiceTitan will happily put two jobs on the board at 8am without understanding that both need the one pump, or that a finisher has to stay on the slab through the finishing window and cannot be two towns over.

Constraint-aware dispatch treats the pump truck, the plant's delivery windows, and each crew's finishing commitments as hard constraints, then routes around drive time between sites. Try to book a pour that would double-book the pump and it stops you before the truck rolls. It sequences the day so the pump runs one sensible loop instead of crisscrossing the metro. Fewer idle crews, fewer frantic 6am reroutes.

The five-star review nobody remembered to ask for

The crew broom-finishes a clean driveway, packs up, and drives to the next job. Nobody asks for a Google review. The competitor two towns over has 240 reviews and shows up first when the next homeowner searches, so you did the better work and lost the ranking. Off-the-shelf tools bury review requests in a checkbox the office manager forgets, and the timing is wrong anyway: asking on pour day, before the slab has cured and the customer has walked on it, gets a shrug.

Automated review requests fire on the right trigger: job marked complete and invoice paid, then a short delay so the customer has lived with the finished work. The text carries a direct link to your Google profile, timed for when they are happiest. An unhappy reply routes to the owner privately first, so a problem becomes a phone call instead of a public one-star. The reviews you were always owed finally show up.

Four years of jobs in ServiceTitan that nobody has mined

Every established concrete company sits on a goldmine it has never opened: years of quotes, completed jobs, and customer records inside ServiceTitan or Jobber, entered by hand and never looked at again. Which job types actually make money after labor and mud. Which estimator wins stamped work and which should stop quoting it. Which customers from three summers ago are due to have their decorative concrete resealed right now. The CRM was built to record that data, not to think about it.

Pointing software at the history turns it into work. A reseal and maintenance engine reads job dates and pings past customers when their surface is due, which is repeat revenue you already earned. Win-rate analysis by estimator and job type sharpens where you spend estimating hours. Costing new bids against what past jobs actually ran, not last year's guess, tightens margin on every quote. The data was always there. Nobody had ever aimed anything at it.

What this costs, and what pushes the price up

Honest numbers, from Digital Heroes delivery experience across more than 2,000 projects. A focused first release, the weather-aware scheduling brain, the AI phone agent, and constraint-aware dispatch, typically runs $50,000 to $120,000 and ships in 10 to 16 weeks. A full operations platform that also mines your CRM history and automates estimates, follow-ups, and reviews runs $150,000 to $350,000, phased over 6 to 12 months so you get working pieces every few weeks rather than waiting a year for a big-bang launch.

What drives the price up in concrete specifically: deep integration into a ready-mix plant's ordering system or a pump scheduling vendor rather than a simple calendar; a phone agent that has to handle both residential homeowners and commercial GCs with different qualifying questions; messy historical data in ServiceTitan or Jobber that needs cleaning before it can be mined; and multi-branch rollouts where each location runs its own crews, equipment, and plant relationships. A single-crew residential shop lands near the floor. A multi-branch commercial operation lands near the top.

When ServiceTitan is enough, and when it is time to build

Straight answer: if you run one or two crews doing predictable residential flatwork, quote at low volume, and mostly need a calendar and clean invoicing, ServiceTitan, Jobber, or Housecall Pro is genuinely enough. Do not spend six figures to replace a tool that is doing its job. The subscription is cheaper than a custom build and it will not fall behind you.

It is time to build, or to layer AI on top, when the signals stack up: multiple crews sharing a pump truck, a schedule you reshuffle by hand every time the forecast moves, after-hours calls going to voicemail, quotes leaking because nobody follows up, and years of data doing nothing. Here is the position we take, though: do not rip out ServiceTitan. It is a fine system of record. Build custom only for the parts it structurally cannot model, the weather-scheduling engine and shared-equipment dispatch, and layer the AI phone agent, follow-up, and review automations on top of it through its API. You keep the CRM your office already knows and add the brain it never had.

How to choose a developer for a concrete contractor

Most software shops will nod along and then build you a prettier calendar. Vet for the specifics of this trade:

• Make them whiteboard a weather-pushed pour. A serious developer should talk fluently about ready-mix delivery windows, pump cancellation minimums, cure windows, and finisher availability. If they treat a pour like a generic appointment, they will build you the tool you already have.
• Ask for proof they have integrated with ServiceTitan, Jobber, or Housecall Pro. The value is layering on top of the CRM your office runs on and migrating years of jobs without losing history. Ask to see a project where they did exactly that.
• Listen to a real call from their AI phone agent. Not a demo script, an actual recording qualifying a concrete job: tear-out, access, square footage. If they cannot play you one, the agent is not real yet.
• Confirm you own the code and the model configuration. You should keep the repository, the data, and the ability to move hosting. If a first release is not shipping in weeks with the rest phased behind it, and if they will not hand over the code, walk.