A building site seen from above, with wall formwork in place and a crane.

Construction

From bid to close-out, with the margin always in sight

FairAI builds AI software for construction companies: quotes from past projects, site start-ups with nothing forgotten, a site diary by message, invoices linked to the right project and actual costs improving the next quote.

Public and private works, main contractors and subcontractors.

Book a 30-minute call

From information to work done

What you already have

ERPERP (PHC…)
SheetsQuotes in spreadsheets
SpecsSpecifications
MessagesSite messages
PhotosPhotos

Work done

Built on what only your company knows

Does this happen in your company?

“Have we built something like this?”Quotes and site diaries from past projects sit in folders nobody can find. See how it is solved See how it is solved
Does every quote start from scratch?And sums already done on other projects get redone. See how it is solved See how it is solved
Material missing on the first day on site?Sheets, panels or signs that did not arrive mean extra trips and lost days. See how it is solved See how it is solved
Does what happens on site stay on WhatsApp?Messages and photos get lost, and rain days go unrecorded. See how it is solved See how it is solved
Do invoices only arrive at month end?And they are typed in by hand, one by one, when it is too late to fix anything. See how it is solved See how it is solved
Is the margin only known at close-out?The gap between quoted and actual shows up when the job is already finished. See how it is solved See how it is solved

Use cases

Where FairAI fits in, from order to delivery

Pick a step: each one has a working example.

Use cases · 1 · Tender

“Have we done this before?” Answered in seconds

1/3 · Ask in everyday words

Archive searchExampleIllustrative exampleIllustrative example with fictitious data

project with synthetic surface and rain during installation

Site diary, 12/11/2025“Work stopped by rain while laying the surface.”Diary · Elmwood School
Quote 25/088Synthetic surface, 340 m², with drainage baseExcel · 2025
Claim CLM 27/011Surface lifting next to the swingPDF · Aftercare
Photo from day 3Described by AI: blue surface, swing areaImage · 2025

4 results, each with its source Use as reference

A technical quote in minutes, not days?

1/4 · The enquiry arrives, with the specification

Quote to clientExampleIllustrative exampleIllustrative example with fictitious data

Enquiry from a client

Brookfield Council

Play area refurbishment, 420 m²

specifications.pdf

3 similar projects in the archive

Urban playground2024 · 380 m²€57,200
School play area2025 · 460 m²€66,900
Garden play space2023 · 300 m²€41,800

Proposed linesSource

Remove existing surface420 m²€2,9402024
Impact-absorbing synthetic surface310 m²€27,900Catalogue
Play equipment4 units€24,600Catalogue
Fencing and access86 m€5,1602025

Margin 20% Total to client €60,600

Reviewed by the estimator Sent

Try it: Exclude or include lines and adjust the margin. The total updates.

Illustrative example with fictitious data

Today

To build a quote, someone searches folders and spreadsheets for similar jobs, redoes the maths and waits for the person who knows the prices to be available.

With FairAI

The enquiry arrives with the drawing. FairAI finds the most similar past jobs and proposes the lines, with the source of every figure. The estimator reviews, adjusts and sends.

With FairAI

Quotes go out even when the most experienced person is away
Prices consistent with what the company has charged before
Every line shows its source, so it is quick to check

Builds on our workthe AI quoting in use at Play Planet, connected to its ERP. Read the case study Play Planet

Talk about this case Next: Nothing forgotten at start-up

Use cases · 3 · Preparation

Nothing forgotten when the project starts

1/4 · The job starts in 7 days

Site preparationExampleIllustrative exampleIllustrative example with fictitious data

67%Project 26/052 · Millbrook ParkStarts 27/10 · 7 days to goPreparation: 4/6

Late

Identification plates: ordered on 13/10, still no delivery date.

Alert sent to the print shop and site management

Try it: Tick or untick items: the preparation and alerts update.

Illustrative example with fictitious data

Today

Site preparation lives in emails and in the site manager’s head. A forgotten plate or board means a special trip and can hold up payment.

With FairAI

Each project has a checklist with owners and dates. FairAI follows up on orders and warns in time about what is still missing, seven days before the start.

With FairAI

An owner and a date for every item
Warnings before a delay becomes a problem
Fewer trips just to bring what was missing

Builds on our worka concept designed with a construction client, on top of the CRM they already use.

Talk about this case Next: Site diary by message

Use cases · 4 · Execution

The site diary writes itself

1/3 · The site group sends voice notes, photos and text

Overview of the site, with the wall formwork in place. — Illustrative example with fictitious data

Concrete mixer truck arriving, next to the excavator. — Illustrative example with fictitious data

Today

What happens on site stays in messages or in the foreman’s memory. When a delay must be justified or a complaint answered, nobody can find the record.

With FairAI

Foremen and workers send voice, text and photos to the project chat. FairAI organises everything and writes the summary of each day: work done, stoppages and issues.

With FairAI

No forms: just speak and take photos
Rain days and stoppages logged
History ready for complaints and extensions of time

Builds on our worka concept designed with a construction client, with the voice and photo reading from the app in use at F. Caixeiro. Read the case study F. Caixeiro

Talk about this case Next: Invoices read and linked to the project

Use cases · 5 · Invoices

Invoices read and linked to the right project

1/4 · The invoice arrives by email

Supplier invoicesExampleIllustrative exampleIllustrative example with fictitious data

Received by email at 17:42

Ridgeway Aggregates LtdVAT no. IE 9999999WINV-1873 20/10/2026

Washed sand 0/4	12 t	€222.00
Gravel 6/14	8 t	€168.00
Tipper truck transport	1	€65.00

Net €455.00 · VAT €104.65Total €559.65

Read automatically

Supplier: Ridgeway Aggregates Ltd
VAT no.: IE 9999999W
Invoice no.: INV-1873
Date: 20/10/2026
Net: €455.00
Total: €559.65

Project suggestedCost category suggested

Posted to the ERP · 26/047 · Elmwood School

Try it: Change the project or category and approve again.

Illustrative example with fictitious data

Today

Invoices for site purchases arrive at the end of the month and someone spends days typing them in by hand, one by one.

With FairAI

The invoice arrives by email or message, is read and gets a suggested project and category. Accounts only approve it, and it goes into the ERP.

With FairAI

Supplier, VAT number, amounts and VAT read automatically
Project cost updated the same day
Nothing goes into the ERP without approval

Builds on our worka concept designed with a construction client, on top of the ERP connection that already reads and writes.

Talk about this case Next: Actuals improve the next quote

Use cases · 6 · Close-out

Actuals improve the next quote

1/4 · The job closes

Project close-outExampleIllustrative exampleIllustrative example with fictitious data

Project 25/088 · Elmwood School · completed

QuotedActual

Materials+8.7%
Labour+15.8%
Equipment0%
Transport-12.9%

Marginexpected 20% → final 14.6%

Next quote · Project 26/061, 400 m²

Materials€39,900
Labour€25,100
Equipment€28,900
Transport€3,200

Total €97,100

Try it: Switch actual figures on or off and watch the next quote change.

Illustrative example with fictitious data

Today

Each project’s margin is only known at close-out, and the same variances repeat on the next project because the quote always starts from the same figures.

With FairAI

At close-out, FairAI compares quoted and actual by category and shows where margin was lost. The next quote starts from the actual figures, and the estimator decides.

With FairAI

Variances by category, project by project
Quotes closer to real cost
Margin protected on the next projects

Builds on our worka concept designed with a construction client, on top of the AI quoting in use.

Talk about this case

Track record

Featured success story: Play Planet

Play Planet builds playgrounds and play areas. It prepares its project quotes with AI, connected to its ERP.

CRM screen with a list of quote items proposed by AI, each with category, unit, price, source and a button to accept it.

In use, connected to their ERP

Playgrounds and play spaces · public and private works

AI quoting in a custom CRM

1Quote lines drafted by AI
2Prices from past work and the catalogue
3Sales accepts or adjusts each line

Connected to their PHC ERP, reading and writing. Next: project management, designed with Play Planet.

Read the case study

More work already done

Document and image search
For a ceramics manufacturer: years of orders in spreadsheets and images, searchable in everyday words.
Search across recorded meetings
For a city council: “Which meetings discussed cycle lanes?” Excerpts, summary, who said it, when, and the minute in the video.
Assistants on documentation
For Pawpass, on the rules for travelling with pets, and for Arbitrum, on the technical documentation for developers.

Chosen by clients, backers and leading programmes

What it costs today

What does repetitive work cost you today?

Run the numbers for your company: looking for information, copying data between systems, putting quotes together by hand.

People doing repetitive work15Hours a week, per person6Cost per hour, fully loaded€18

What your team already spends today

392 h a month

In wages, this already costs you about €7,047 a month (€84,564 a year).

The maths: 15 people × 6 h a week × 4.35 weeks × €18 an hour.

In the call, we look at where that time weighs most and what can be done with it.

Book a 30-minute call

FAQ

What people usually ask us

Do we have to change our ERP?

No. The ERP (PHC, for example) stays the system of record, and nothing goes into the accounts without approval, unless agreed otherwise. FairAI connects to it and to what comes in from site.

We already have AI in our ERP, or Copilot. Why do we need you?

Use it for what it does well. Copilot and the AI built into ERPs (SAP, Microsoft Dynamics, Sage…) help with emails, summaries, finance and alerts, and they are the same for every company. We take care of what is yours alone: the invoice linked to the site and cost centre, the site diary, the next estimate built from actual costs. We connect to the ERP; we don’t replace it.

Do site managers have to learn a new app?

No. The site diary uses what they already use: voice notes, text and photos in the site group chat.

Have you done this at a construction company?

Yes. AI quoting is in use at Play Planet, connected to its ERP, and we are designing site management with them.

Who approves invoices before they go into the ERP?

Accounts. FairAI reads the invoice and suggests the project and cost category, but nothing goes into the ERP without approval.

Does our data leave the company?

It does not have to. The AI can run on your own servers, with local models, without FairAI having access to the data. If you prefer the cloud, which costs less, we put in writing what leaves and where it goes.

Not sure where to start? Take a 2-minute assessment

From bid to close-out, with the margin always in sight

Does this happen in your company?

Where FairAI fits in, from order to delivery

Featured success story: Play Planet

AI quoting in a custom CRM

More work already done

What does repetitive work cost you today?

What people usually ask us

Let’s look at a process in your projects