Svennis AI
Artificial intelligence · AI for builders

Artificial intelligence (AI) for builders: how to run a tighter building firm

See how AI can be applied to the real processes of a UK building firm: enquiries, priced quotes and take-offs, the site diary, subcontractor payments, stage payments and job margins. SME builders reported net enquiries up 18% in the second half of 2025, yet more than half still reported lower profits than they had expected, or losses, so the firm that quotes fast and controls its costs keeps the work. Each process comes with examples and an honest view of the technology.

Pick an area of your business

Enquiries and quotes: a priced answer while the job is still warm, not next week

For most building firms the work is there; the trouble is turning it into a priced quote fast enough. SME builders logged net enquiries up 18% in the second half of 2025, on top of a 34% jump in the first half, so the pipeline is not the constraint. The constraint is that enquiries land while you are on the tools, up a scaffold or mid pour, and quoting gets pushed to the evening. The slowest quote usually loses the job.

A quote worth defending is built from measured quantities, a take-off, not a lump sum plucked out of the air. The RICS New Rules of Measurement (NRM2) are the UK standard for that build-up, and a firm that prices from measured work can stand behind its number, see what a rival missed, and reuse the same measure downstream in the valuations.

There is a moving target underneath every fixed price. Building materials rose 3.0% in the year to November 2025 in aggregate, but the aggregate hides the risk: imported sawn wood and plywood were up 11.9% while steel reinforcing bars fell 7.0%. A quote priced a few weeks ago can be materially wrong by the time the order goes in.

An assistant we build for your firm is built for precisely that gap. It reads the incoming enquiry, asks the two or three qualifying questions a good estimator would ask, drafts a structured brief for pricing, and flags which material lines are exposed to movement. It never sends a price on its own. The commercial call, hold the number or go back to the client, stays with you.

WhatsApp· AI assistant
Hi, we're after a quote for a single-storey rear extension, about 25 square metres, kitchen-diner, with a bifold to the garden. Roughly what would that cost and how soon could you come and take a look?
Good evening, and thanks for getting in touch. I'm the firm's AI assistant, so I can get the details down and book you a visit, but the price itself will come from our estimator after we've seen the site. A proper number on an extension has to come off measured work, not a guess. Could I ask a couple of things first: do you already have drawings and planning or building control approval in place, and roughly when are you hoping to start?
We've got architect's drawings and planning is approved. We'd like to start in the spring if we can.
That's really helpful, and having approved drawings makes the take-off far more accurate. I can book one of our team to survey and measure up. We have Wednesday afternoon or Friday morning free this week. If you can have the drawings to hand on the day, we'll get the quantities off them on site.
Friday morning works. Can you give me even a rough figure now, so I know we're in the right range?
I'd rather not put a figure in front of you that I can't stand behind, because the groundworks and the steel over the bifold can swing the number a fair bit. What I'll do is book Friday at 9:30, and our estimator will send a written, itemised quote after the visit, with the assumptions set out so you can see what's in it. Does 9:30 suit?
01

A first reply to every enquiry, day or night, off the site and off your evening

The assistant is connected to your enquiry inbox, your website form and your WhatsApp. When an enquiry arrives it responds straight away: it thanks the client, confirms the firm covers the work and the area, and asks the handful of qualifying questions that decide whether a job is worth a site visit. It introduces itself as an AI assistant and books the survey into your diary rather than leaving the enquiry to cool in an inbox.

Example

A homeowner in Reading messages at nine in the evening about a single-storey rear extension, roughly 25 square metres, kitchen-diner, bifold to the garden. The assistant replies within minutes, confirms you take on extensions of that size, asks whether drawings and planning are in place, and offers two survey slots for the week. You open a booked appointment in the morning, not a two-day-old message.

The benefit

No enquiry sits unanswered while you are on site, and the client hears back from you first. Your evenings stop being the only time quoting happens, because the qualifying conversation is already done by the time you sit down.

02

A structured brief drafted for pricing, so the take-off does not start from a blank sheet

From the enquiry and the survey notes the assistant assembles a first-pass brief in the shape your estimator works in: the elements of the job, rough quantities, the assumptions it has had to make, and the questions still open. It works strictly from what it has been told and marks anything it has assumed, so a person prices real measured work rather than reconstructing the job from a chat thread.

Example

After the survey on the Reading extension, the brief lists the substructure, the frame, the roof, the bifold opening and the kitchen fit, with provisional quantities and a note that the ground conditions and the beam design still need the structural engineer. The estimator prices from a structured starting point instead of re-reading the whole enquiry.

The benefit

The take-off begins half-built and consistent from job to job. Less time is lost re-keying details the client already gave you, and the quote goes out while the client is still comparing firms.

03

Exposed material lines flagged before the price commits the firm

Because prices move by item, not evenly, the assistant checks the material lines in a draft quote against current reference rates and flags the ones that have shifted enough to matter. It does not change the price. It hands you a short list: these lines are exposed, this is the movement, confirm with the merchant before you commit.

Example

A quote drafted a fortnight ago for a Taunton loft conversion carries a large timber allowance. The assistant flags that imported sawn wood and plywood have moved sharply since the price was set, and suggests confirming the joist and sheet-material lines with the supplier before the quote is sent. You re-price two lines rather than re-reading the whole sheet.

The benefit

The repetitive number-watching that erodes fixed-price quotes is done for you. A job is far less likely to slide from profit to loss between tender and build because a volatile line went unchecked.

How ready the AI technology is

The pieces here are buildable today, and the market data shows where the payback sits.

  • The demand signal is measured, not assumed: SME builders reported net enquiries up 18% in the second half of 2025, and up 34% in the first half, which tells you the bottleneck for most firms is conversion speed, not lead volume. An assistant that reads an enquiry, asks the qualifying questions and books the survey plugs straight into your inbox, website form and WhatsApp, and that integration can be built now.FMB
  • The pricing discipline it supports is a recognised UK standard. RICS NRM2 sets out how a bill of quantities is prepared to obtain a tender price, so an estimate can rest on measured quantities rather than a round figure. An assistant can structure the first-pass take-off from the drawings and the survey notes, far faster than by hand, and leave the quantities and rates for a person to correct and price.RICS
  • The material-exposure check runs on real published data. Building materials moved 3.0% over the year to November 2025 in aggregate, but with wide item-level swings, imported sawn wood and plywood up 11.9% while steel reinforcing bars fell 7.0%. Flagging which lines of a fixed-price quote are exposed is exactly the repetitive number work an assistant can prepare for a human to sign off.GOV.UK / DBT
  • Keep your expectations anchored to your own figures. The numbers worth trusting are the time from enquiry to quote sent, the quote-to-win rate, and how many quotes leave the office in a week, each measured before and after. Any productivity percentage a tool vendor quotes is drawn from other firms and other markets, so read it as a direction of travel, not a promise for your books.FMB
What to watch out for

Three points belong in the build from the start.

  • A quote drafted or priced with help from AI still binds the firm the moment it goes to the client. Quantities, rates and assumptions need a competent human check before sending, because an error in the take-off carries straight through to the interim valuations and the final account. The assistant prepares; a person prices and signs off.RICS
  • There is no UK AI statute and no blanket duty to disclose AI use, but the rules that do exist are technology-neutral: the ASA and CAP Codes apply however content is generated, and a misleading claim is misleading whether a person or an assistant wrote it. In practice that means the assistant quotes only what the firm can actually deliver, and says plainly that it is an AI assistant.ASA
  • Enquiry details, the name, address, property, drawings and photos, are personal data under the UK GDPR as amended by the Data (Use and Access) Act 2025. Because an external AI tool that touches them is your processor, the build has to carry a signed processor contract, a tight retention window and only the fields the pricing actually needs. Get it wrong and the ceiling is a fine of up to £17.5 million or 4% of worldwide turnover.ICO
Svennis Cloud Solutions

We build them, on Claude

These AI flows do not stay on paper. Svennis Cloud Solutions builds and integrates them into your systems, with a team of certified Claude architects, on Anthropic technology, from the first WhatsApp message to the finished invoice.

See live what we can doCertified Claude architects

See what this looks like in your building firm

Describe your building firm in one sentence and Claude will show you live where AI would make the biggest difference, using your own numbers.

Guide: AI for any company
The processes every business has, in detail.
AI law for business in the UK
The rules that govern AI use in UK business, regulator by regulator, with the fines that apply and links to the official sources.
Sources
  1. 1. FMB - State of Trade Survey H2 2025
  2. 2. RICS - New Rules of Measurement (NRM)
  3. 3. GOV.UK / DBT - Construction building materials commentary (data to November 2025)
  4. 4. ASA - Disclosure of AI in advertising: striking the balance between creativity and responsibility
  5. 5. ICO - Guidance on AI and data protection
  6. 6. legislation.gov.uk - Building Safety Act 2022
  7. 7. CITB - Construction Workforce Outlook 2025-29
  8. 8. GOV.UK - CIS: make deductions and pay subcontractors
  9. 9. GOV.UK - VAT domestic reverse charge for building and construction services
  10. 10. The Joint Contracts Tribunal (JCT) - standard building contracts
  11. 11. legislation.gov.uk - Housing Grants, Construction and Regeneration Act 1996
  12. 12. GOV.UK - Making Tax Digital for Income Tax (sole traders and landlords)
  13. 13. GOV.UK - Promoting electronic invoicing: consultation response
  14. 14. ICO - Monitoring workers guidance
  15. 15. legislation.gov.uk - Equality Act 2010
  16. 16. legislation.gov.uk - Data (Use and Access) Act 2025, section 80