We go through how the business really runs: the enquiries you never priced, the variations nobody wrote down, the RAMS reissued for the fourth time. Then we tell you which of it is worth handing to AI, and in what order.
Most contractors know AI could take some of this off them. What they are missing is someone who will look at their own jobs and say which part to hand over first.
A client asks for something extra, the foreman says yes, the work gets done. Six months later at final account there is nothing in writing and you are arguing about whether it was in the contract.
There is an AI feature in the takeoff software now. The trade press is full of it. None of them tell you which job in your week they would actually take off you, and most of them assume you are a main contractor with a document controller.
Every enquiry gets a proper takeoff whether it was ever going to happen or not, because there is no quick way to tell the serious ones from the price-checkers. That is nine estimates you paid for and nobody bought.
None of that is a technology problem. It is a question of which job goes first, and that is the bit we do.
These are the ones that come up in nearly every construction audit we run. Which of them lands in your plan, and in what order, depends on the work you take, your team, and the systems you already run.
Every enquiry gets a takeoff, including the ones that were only ever collecting a third price.
Asks the questions you would have asked about budget, timescale, planning status and decision maker, and sorts the enquiries worth an estimator's day from the ones that are not.
Whether to price it. The automation gathers the facts, it does not decide what you bid for.
The last similar quote opened up, edited, and the previous client's address found in it halfway down page two.
Builds the quote from your own rates and past jobs of the same type, in your document format, with the exclusions and qualifications you always mean to include.
The rates, the margin and the risk allowance. Pricing a job is the commercial judgement and it stays yours.
Changes agreed verbally on site, remembered differently by both sides at final account.
Turns a site note, voice message or photo into a written variation, prices it against your rates, and sends it for the client's confirmation while everyone still agrees what was said.
The price and whether it is a variation at all. The QS or the owner signs it off before it goes.
A method statement adapted from the last job, and a risk assessment that still names a site you finished in spring.
Drafts the RAMS for the job from your own library against the work and the hazards actually present, and flags what needs a competent person to review.
A competent person reviews and signs every RAMS. Nothing here is a substitute for that and the plan says so in writing.
Insurance certificates, qualifications, UTR and CIS verification chased by email every time a new subbie starts.
Requests the documents, checks what has expired, chases what is missing, and keeps the compliance pack current across every subbie on the books.
Whether the subcontractor is any good, and the CIS deduction treatment where it is not clear cut.
The monthly valuation built by hand from site records, submitted late, and paid later.
Assembles the application from measured work, variations and materials on site, to the deadline in the contract, and chases the payment notice when it does not arrive.
What you claim for. Over-claiming is a commercial decision with consequences and it stays with a person.
Clients ring for an update because nobody has told them anything since the slab went in.
Drafts the weekly update from site photos, the programme and the day's records, so the client hears from you before they chase.
Anything about delay, cost or a problem. That is a conversation, not a notification.
A walk round with a notebook, then an evening typing it up and working out who is fixing what.
Turns the notes and photos into a structured snagging list, groups the items by trade, issues them to the right subbie, and chases until they are signed off.
What counts as a defect. Somebody who knows the job walks it.
Written up from memory on Friday, or not at all, which is a problem the first time there is a delay claim.
Assembles the daily record from labour on site, deliveries, weather and photos, so the contemporaneous evidence exists without anyone having to remember to write it.
The site manager confirms it. An unchecked record is not evidence.
Somebody rings three suppliers to find out where the windows are, usually the morning they were meant to arrive.
Tracks what is ordered against the programme, chases confirmations, and raises the delivery that is going to be late before it holds the trades up.
Choosing suppliers and accepting a substitution.
Certificates, warranties and datasheets collected from everyone at the end, weeks after practical completion, delaying the final payment.
Collects each document as the work happens, chases what is outstanding, and assembles the pack in the structure the client wants it in.
The final check that the pack is complete before it is issued.
Nobody starts with eleven. A plan usually puts two live in the first month and leaves the rest ranked and waiting, so when you have the appetite for the next one you already know which it is.
Most of the job is joining up the systems you pay for rather than moving you off them. These are the ones we come across most in construction, and we have no affiliation with any of them.
Three steps. You do the talking on the first one, and nothing is committed until you have seen what we found.
No prep, no agenda to fill in. We ask the questions and you walk us through how your week actually runs: the jobs, the tools, the emails you send a dozen times a month.
Recorded and written up, so you never repeat yourself.
We work back through the call and map where your time actually goes. Then we find the automations that suit your tools, your team and your budget, ordered so the biggest time saving comes first.
Back with you within a week.
We walk you through the whole report, recommendation by recommendation, and you decide what to take on first. Most people have their first automation running within a week of that call.
The report is yours to keep either way.
The plan tells you what to build. The dashboard shows you what it is doing once it is built, so you can see what you are getting back without asking us.
Set up when the plan is delivered, at app.fewerhours.com under your business name. No shared login and nothing to install.
Every one we build reports the hours it saved and the work it handled, so the value is a number you can check rather than a promise we made.
The rest of the ranked list sits there waiting. When you have the appetite for the next one, you already know which it is.
This works best where the owner still knows every job on the board. There are a few cases where it would waste your time, and those are worth saying out loud.
Fewer Hours exists because too many owners knew AI could help them somewhere and had no way of telling a good idea from a sales pitch. The plan is the boring part done properly: sitting with the real workflow and working out which jobs are worth handing over.
Whoever takes your call is the person who writes your plan. No account manager, no junior doing the analysis, and nobody trying to move you onto a retainer at the end of it.
Three steps, and you can stop after any of them. The first one is a conversation, and nothing about it commits you to the next.
A short chat about how your business runs. If there’s nothing here worth mapping, we’ll say so on the call.
The full job. We map the workflows, score every candidate, and hand over the ranked list and the 30-day plan.
If you’d rather we built the projects than just planned them, we’ll do the ones the audit put first. Anything from a quick win to a full agent build.
If there are fewer than five of you and admin takes about an hour a day, there probably isn’t enough here to be worth mapping, and we’ll say so. If you already know exactly which job you want automated, skip the audit and just ask us to build it. And if your processes only exist in your head, we’ll tell you to go and write them down first.
If yours is not here, ask on the call. It is twenty minutes about how the business runs, not a pitch.
It can draft them from your own library and the hazards actually present on the job, which removes most of the typing. It cannot sign them. A competent person reviews and approves every one, the legal responsibility stays with the firm, and any plan we write says that explicitly rather than leaving it implied.
It is usually an opportunity. Voice notes and photos are exactly where variations and site records get lost, and turning them into written, dated records is one of the highest-value things on most construction plans. You do not have to change how the site communicates for that to work.
Every recommendation states what data it touches and where that data ends up. Keeping contract and pricing data inside systems you already control is the default, and anything that would move it somewhere new is flagged, with the alternative, before it reaches the plan.
That is one of the first things we check. Some of it is the API, some is a feature your package already ships and nobody switched on, and some is email and a shared folder joined up properly. Where your system will not do something, the plan says so rather than pretending.
It assembles the quote from your rates and your past jobs, which is most of the typing. It does not set the margin or the risk allowance, because pricing is the commercial judgement that decides whether you make money and it should never be automatic.
Then the plan should not need them to, and most of ours do not. The highest-value automations sit in the office, working from things site already produces: photos, voice notes, delivery tickets. If a recommendation depends on changing site behaviour, we say so, because that is the thing most likely to fail.
The plan lands about two weeks after the call. Most contractors have the first automation live within a week of that, because the first one is usually enquiry qualification or variation capture and neither needs anyone to change system.
Either. The plan is yours to take to anyone, including whoever supports your accounts package. If you would rather we set up the first two, that is a separate job and you decide after you have read the plan.
The call takes twenty minutes and it is about how the business runs, not about software. The plan lands about two weeks later.