The solution
In the field

Mobile app

The CPQ in the salesperson's pocket. A voice note, a few photos, and the customer walks away with a price range before you have left the site.

What gets lost between the visit and the quote

A site visit produces a great deal of information: measurements taken by hand, a nameplate photographed, an access constraint mentioned in passing, the budget the customer eventually lets slip. Today all of it travels through a notebook, a few photos in the phone gallery and the salesperson's memory.

Re-keying happens that evening, or the next day, or the following week. At each step, substance disappears. The quote ends up going out several days after the moment the customer was most ready to buy, and it is usually that delay that costs the deal.

Three gestures during the visit

The app doesn't ask you to fill in a form in the middle of a workshop. It takes what the field naturally offers, speech and images, and turns that into usable parameters.

Dictate the survey

Recording · 0:42

“We're on a 120-metre network, four drop-points…

…the customer wants 45 cubic metres an hour at peak.”

Smart Tags detected

Flow 45 m³/hLength 120 m4 drop-points

01

Dictate the survey

The salesperson speaks, DARYL transcribes and extracts parameters as the sentence unfolds.

Photograph what's there
DN 80
PN 16
Nameplate read

Smart Tags detected

DN 80 — PN 16Galvanised steel

02

Photograph what's there

Nameplate, pipework, a drawing pinned to the wall: DARYL reads the references, diameters and materials off it.

Price on the spot

Budget estimate

Martin & Co — Langon site

  • Equipment€14,200
  • Installation€5,400
  • Commissioning€1,800

Budget estimate

€21,400 – €23,900

Justified · traceable

Send to customer

03

Price on the spot

A justified budget range, handed to the customer before leaving the site.

  • Hands-free voice note

    The salesperson comments on what they see while they see it, without putting down the tape measure. DARYL transcribes and, above all, interprets: “120 metres of network, four drop-points, 45 cubic metres an hour at peak” becomes three populated Smart Tags, not a paragraph to re-read later.

  • Photos become data

    Nameplate, fitting, a drawing pinned to the wall, a sketch on a scrap of paper: the photo is analysed by the multimodal engine. References, diameters, materials and standards are extracted and attached to the right customer record, with the image kept as supporting evidence.

  • Smart Tags editable on the spot

    Every extracted parameter stays visible and editable with one tap. If DARYL misheard a dimension, the salesperson corrects it in front of the customer, and that correction feeds the learning loop.

A first estimate before you get back in the car

Once the parameters are captured, the sizing engine runs on the very same domain rules as the desktop version. The salesperson gets a justified budget range: every line traces back to a rule or to a comparable past job, and stays open to inspection.

The offer goes out the same day

The customer receives an estimate while still inside their project, not three weeks later when they are already comparing three competitors.

No signal, no problem

Plant basement, warehouse, farm plot: capture works offline and syncs as soon as connectivity is back.

One record, two screens

What is captured on mobile is already in place at the desk: the record opens on the workstation with its Smart Tags, media and pre-sizing.

In the field, concretely

Agricultural irrigation

The sales engineer walks the plot, dictates lengths and elevation changes, photographs the existing pumping station, and hands over a network estimate before driving off.

Industrial acoustics

Survey of noise sources on the shop floor: photos of the machines, a voice note on access and operating constraints, pre-sizing of the treatment right after.

Air treatment and fluid networks

Reading the nameplates of compressors and air-handling units, rebuilding the existing network from the photos, pricing the replacement on site.

Why this estimate holds up

Producing a price in three minutes only matters if the price holds. That is the job of the engine underneath: an LLM for language, a rule graph for the decision.

See the AI technology