Construction automation: quotes, tenders, progress billing

Checking subcontractor quotes, filling in prime contractors' quote formats, screening tenders, and matching progress billing, with AI and human review.

  • Runs on your own PCs and servers
  • Use Azure OpenAI, Amazon Bedrock, Google Vertex AI, or an in-house model
  • Its engine is Descarty's open-source Dagu (GPL-3.0)

Common problems

  1. Since December 2025, quotes are expected to itemize labor and other costs, and each subcontractor itemizes in its own way

  2. The same breakdown is copied again into each prime contractor's own quote format or quoting system

  3. Overtime caps have applied to construction since April 2024, while the workforce keeps shrinking and ageing

Construction paperwork moves between prime contractors and subcontractors

Japan has 483,823 licensed construction businesses (end of March 2026), and 98.9% of them are sole proprietors or companies with capital under 100 million yen. The industry employs 4.77 million people (2024), of whom 36.7% are 55 or older and 11.7% are 29 or younger.

Work passes from the prime contractor to first-tier, second-tier, and further subcontractors. Quotes, orders, safety documents, and invoices move back and forth between them.

The industry already has systems for much of this. Safety documents (known as the “green file”) can be created and submitted through dedicated services such as Green Site and Buildee. Green Site has more than 130,000 contracted companies (end of August 2026), and it generates the site’s subcontracting structure register and warns of expired health checks, licenses, and construction permits. The days each worker spends on each site are recorded in the Construction Career Up System (CCUS), the national register of construction workers’ skills and work history, where 1.92 million workers and 317,000 businesses are registered (end of September 2026). Site entry services such as Green Site, Buildee, and ANDPAD connect to CCUS through its API. For exchanging data from quote requests through orders and invoices, the industry has a standard called CI-NET.

Work What already exists What stays manual
Checking tender notices The government procurement portal’s API Reading notices and picking the ones you can bid on
Bidding on public works Each agency’s electronic bidding system Preparing the bid price breakdown (labor and other costs required since December 2025)
Checking quotes (prime contractor) CI-NET Checking the breakdown in quotes that arrive as PDF or Excel files
Preparing quotes (subcontractor) CI-NET Copying the breakdown into each prime contractor’s format or quoting system
Safety documents Green Site, Buildee, and similar services Upper-tier companies preparing and submitting documents for subcontractors that can’t use the service
Workers’ site work records CCUS and site entry services connected through its API, with automatic transfer to the construction workers’ retirement fund Registering the site structure in CCUS when the service used doesn’t send it
Progress and invoices CI-NET Matching orders, assessed progress, and invoices

From December 2025, quotes show labor costs

Japan’s amended Construction Business Act, promulgated in June 2024, took effect in stages in September and December 2024 and fully on December 12, 2025. One aim is to secure labor costs, which fund workers’ wages, all the way down to subcontractors. The Ministry of Land, Infrastructure, Transport and Tourism (MLIT) explains that under the custom of lump-sum contracts, the going rate for labor is hard to see and labor is easier to cut than materials.

On December 2, 2025, the Central Construction Industry Council recommended a labor cost standard. A fair labor cost is the public works design labor rate multiplied by a productivity rate (worker-days per unit of work). MLIT publishes reference values for each trade as labor cost per square meter, per ton, and so on. The design labor rate is set for each of 47 prefectures and 51 trades, and the rates in effect from March 2026 rose 4.5% from the previous year on a simple national average.

The main rules after the amendment:

  • Contractors should try to issue quotes that itemize materials, labor, statutory welfare costs, health and safety costs, and retirement fund contributions, and those ordering the work should try to take them into account
  • Contractors must not quote with markedly low labor costs, and those ordering must not ask for a quote to be changed that way
  • The ban on contracts below cost now applies to contractors as well
  • Bids for public works must come with a price breakdown that shows labor and other costs

In fiscal 2025, MLIT’s investigators examined 1,152 businesses and gave guidance or advice to 768. The most common issue was missing itemization or unclear conditions in quotes, with 641 cases (including guidance on provisions not yet in force). A January 2026 collection of problem cases lists practices such as leaving labor rates unchanged for years, and using the lowest of several quotes to push other bidders down. Negotiating a discount is not ruled out, but MLIT says any discount has to come out of the contractor’s profit margin, not labor costs.

There are as many quote formats as business partners

Itemized quotes take effort both to prepare and to check. In March 2026, MLIT published an operational guide built on issues raised in a pilot project. Its Q&A lists, from the subcontractor side, that preparing itemized quotes adds a heavy workload, and from the prime contractor side, that checking is cumbersome because each subcontractor accounts for health and safety costs differently.

Formats are not standardized. In an MLIT survey, about half of specialist trade associations had prepared a standard quote template or were working on one. MLIT’s operating policy allows the party ordering the work to specify the quote format. A participant in the pilot noted that, as a first-tier subcontractor, it will still have to prepare quotes in each prime contractor’s own format.

Quotes also have to be kept. Construction businesses must keep the initial and final quotes for 10 years from handover of the completed work, and the operating policy recommends keeping them electronically.

Where both sides use CI-NET, quotes travel as data. Where quotes go back and forth as PDF and Excel files or through each prime contractor’s quoting system, reading, copying, and comparing stay with people.

Overtime caps now apply to construction

Since April 2024, overtime caps have applied to construction. The basic limit is 45 hours a month and 360 hours a year. Even under a special clause, overtime is capped at 720 hours a year, under 100 hours in any month, and an average of 80 hours over several months, with holiday work counted in the last two. Disaster recovery and reconstruction work has exceptions.

Construction workers put in 2,018 hours a year, against 1,956 hours across all industries (fiscal 2023). When the Ministry of Health, Labour and Welfare inspected workplaces suspected of long working hours in fiscal 2025, it found legal violations at 1,878 of 2,421 construction workplaces, and illegal overtime at 865. In one published case, an electrical contractor had workers with up to 237 hours of overtime and holiday work in a single month.

With fewer workers and capped overtime, preparing and checking itemized quotes is new work that has to fit in.

What to check before automating

  • Whether a dedicated service is enough: Green Site and Buildee for safety documents, services connected to CCUS for site work records, and CI-NET for exchanging quotes through invoices. If they cover what you need, use them first
  • Whether you may operate a partner’s system: check the terms of use and your trading terms for each prime contractor’s quoting system and each agency’s site
  • Who decides: staff decide labor rates, productivity rates, quote amounts, revision requests to subcontractors, and which company gets the contract. AI reads, compares, and drafts
  • What to compare against: the design labor rate is revised every year, and the government’s reference values assume standard site conditions. Decide first which year’s rates, and which prefecture and trade, to compare with
  • Where quotes are kept: initial and final quotes must be kept for 10 years from handover. Decide where they go and how they link to each project
  • Submitting electronic bids: bids are submitted with an IC card issued by a certification authority. Staff submit them

What Kitewell does

Descarty’s Kitewell is integration and automation middleware that connects systems through APIs, files, databases, and screens.

  • Reads and compares quotes: questions to AI and judgments can be workflow steps. AI reads the breakdown from quotes in PDF, Excel, or image form, compares it with your rate tables, and drafts points to confirm. The AI model is chosen per step, including in-house and local models
  • Operates prime contractors’ quoting systems: AI reads and operates web systems and desktop apps that have no API. Operations that worked are remembered and repeated the same way on later runs
  • Keeps running when screens change: when a remembered operation no longer fits the screen, AI reads the screen again, does the step, and remembers the new operation. If that still fails, the run stops, and AI looks into the cause and suggests the fix
  • Starts when email arrives: Gmail, Microsoft 365, and IMAP mailboxes are checked at set intervals, and attached quote requests and quotes are saved
  • Handles sign-in hurdles: when a one-time password is required, the run pauses until staff enter it. Passwords are referenced by name and typed in without showing the value to the AI
  • People check where it matters: approval can be set on any step. Requests reach Slack, Microsoft Teams, and email, and sending one back restarts the run from the chosen step
  • Every run on record: each run keeps its logs, screenshots, and approval decisions, so you can later see how each quote was read and who approved it
  • Keeps each prime contractor’s rules: knowledge stores how to fill in each prime contractor’s format and its exceptions, per project, and the ledger of workflows can be exported

You can start from one piece of work, such as preparing quotes in each prime contractor’s format through submitting them and recording won jobs in your ledger. A demo shows how it runs, a free connection test confirms Kitewell reaches your systems, and a one-month PoC measures the result against success criteria agreed up front, before you license it for production. See From evaluation to production for details.

Sources

Frequently asked questions

Does Kitewell judge whether a quote's labor cost is adequate?
No. AI reads the breakdown and points out missing cost items and labor rates below the public works design labor rate. The government's reference values assume standard site conditions. Staff judge whether productivity rates and costs fit the actual site, and whether to ask the subcontractor for a revision.
We use Green Site or Buildee. What would Kitewell do?
Dedicated services like these already handle creating and submitting safety documents, flagging expired items, and sending work records to CCUS. Keep using them. Kitewell suits work such as quotes, where every business partner uses its own format and no dedicated service exists.
Can electronic bidding be automated too?
Bids are submitted with an IC card issued by a certification authority and a card reader. Staff submit bids, including the decision to bid at all. What can be automated is the work before that: collecting tender notices and picking the ones you are eligible for.

Use cases for any industry

  • All industries
  • Accounting & finance

Read invoices into the accounting system

AI reads incoming invoices, checks them against purchase orders, and records them in the accounting system. Invoices whose amounts differ wait for approval.

  • All industries
  • Accounting & finance

Create invoices and send them by email

Invoices are filled into your template from an Excel billing list and emailed once a person approves. AI first points out amounts far from last month's.

  • All industries
  • Sales

Research a list of prospects

For an Excel list of hundreds of companies, each one's business, locations, and size are collected from its website into its row, with a priority for sales.

  • All industries
  • HR & payroll

Find missing clock-ins

Collect yesterday's clock-ins from the attendance system and, once HR approves, ask employees with missing entries to fix them in Teams.

  • All industries
  • Accounting & finance

Take in emailed invoices and draft the entries

Find unread email with an invoice attached, have AI read the amount, and record a draft ledger entry. Uncertain entries wait for accounting to confirm.

  • All industries
  • IT

Review SaaS accounts

A SaaS account review matches each service's users to current staff. Collect each console's users, match them, and disable leavers only after approval.

  • All industries
  • Accounting & finance

Partner registration that starts on approval

A kintone approval starts the workflow by webhook. It looks up the tax registration number, AI matches it to the request, and a match enters the sales system.

  • All industries
  • Planning & marketing

Collect KPIs from SaaS tools into a weekly report

Collect figures from advertising, sales, support, and other SaaS tools, and produce a weekly report with week-on-week comparisons and likely causes of change.

  • All industries
  • Accounting & finance

Check supplier invoice registration numbers

Check your supplier master's registration numbers monthly through the NTA's Web-API, and list suppliers that were revoked, lapsed, or differ in name.

Prove Kitewell on your own systems

Try it free for 14 days on your own PC. To see it working first, ask for a 30-minute demo. There is also a one-month onboarding (¥500,000 before tax), where we build your first job together.

What the onboarding checks

  • Which method connects each target system: API, file, database, or screen
  • Whether Kitewell can work a system with no API through its screens, as a person would
  • Whether AI relearns a step and processing continues when a screen changes
  • Whether errors are handled as agreed: stop, notify, or retry
  • Whether it runs in your environment, with the AI model you choose
  • Whether processing time and AI usage cost fit daily operation

How the onboarding runs

  1. Week 1Requirements and success criteria
  2. Weeks 2 to 3Build
  3. Week 4Verify on production screens
  4. Final dayReport

One-month onboarding

Length
One month
Cost
¥500,000 (tax excluded). Your staff build, with the developer beside them
If you go on to a license
The onboarding fee is credited to the first year's contract
Without the onboarding
At Team scale, start on Team monthly (¥14,800 per computer a month, tax excluded), and cancel in any month

You can try it first on your own PC, in the free 14-day trial.

See how the onboarding runs and its terms