Landscape takeoff that
reads the L-sheets.

The plan and the schedule never quite agree

Every landscape estimator knows the drill. The planting plan shows symbols scattered across the beds; the planting schedule lists quantities, sizes and spacing in a table on a different sheet; the hardscape sections live in the details; and the irrigation runs cross the whole thing on its own L-sheet. You're not measuring one drawing — you're cross-reading four, and the bid only resolves when they reconcile. That's slow work by hand, and it's exactly where things slip: the schedule says 142 shrubs, you count 137 symbols, and under deadline you just trust the table and move on.

Pilars reads the planting schedule as a structured table, counts the symbols on every L-sheet, and reconciles the two. Where they disagree, you get a flag and a location instead of a buried assumption. The same pass classifies the bed hatches, turf areas and hardscape zones against the legend, then pulls the depths and section details that turn an area into a quantity. You review a confidence-scored list against the plan rather than highlighting beds one drawing at a time.

Where landscape bids actually leak

It's almost never the plant count itself — it's everything stacked around it. The plant schedule count disagrees with the symbols on the plan and nobody catches which is right. The mulch is specified at 3 inches and the topsoil amendment at 6 inches, but those depths are buried in a paragraph of planting notes, so the bed cubic yards get estimated flat. Sod, hydroseed and seed read as one lawn area because separating the hatches is tedious. The steel edging linear footage gets eyeballed instead of traced. And the hardscape base — the part that actually costs money — comes off a default depth because nobody opened the paver section to read the 6-inch aggregate base and the sand setting bed.

Pilars reads off the L-sheets the way you would if you had unlimited time. It reconciles schedule quantities against plotted symbols, pulls mulch and topsoil depths from the notes and applies them by bed area, splits turf into sod versus seed versus hydroseed, traces edging linear footage, and reads the hardscape section details for base course depth, setting bed and geotextile. Every quantity is tied back to the line on the sheet it came from, so when a number looks off you can see why in one click.

• Plant counts reconciled: schedule quantities vs plotted symbols, with discrepancy flags
• Mulch and topsoil depths read from planting notes, applied as CY by bed
• Sod vs hydroseed vs seed areas split from the turf hatches
• Steel, aluminum and concrete edging traced as linear feet by type
• Hardscape base sections read for aggregate depth, sand setting bed and geotextile

What the AI actually reads off an L-sheet

Landscape drawings are dense and inconsistent across firms, so the reading matters more than the measuring. Pilars starts with the legend and the planting schedule, builds a mental model of what each symbol and hatch means on this particular set, and then applies it. A 24-inch box tree, a 5-gallon shrub and a 1-gallon perennial are different costs and different counts, and they're distinguished by the size column in the schedule, not by how the symbol looks. Spacing notes — "shrubs at 30 inches on center" — turn a bed area into a real plant count where the designer drew a fill instead of individual symbols.

On the hardscape side it reads the section details that estimators routinely skip under time pressure: the depth of compacted aggregate base under pavers, the bedding sand, the geotextile separation fabric, the edge restraint. On the soft side it reads the depths and amendments that drive the dirt work — topsoil import, soil prep, mulch. The output is a takeoff organized the way a landscape proposal is organized: plant material by size, turf by method, beds by depth, hardscape by assembly, edging by type and irrigation handed off cleanly to the irrigation scope.

Honest about what it is

Pilars reads 2D PDF site and landscape plans — the same L-sheets and details you bid from today. It is not a magic site model and it does not invent quantities that aren't on the drawings; when the planting schedule and the symbols disagree, it tells you they disagree rather than picking a winner. You stay the estimator. What changes is that the tedious cross-reading between the schedule, the notes and the details happens in minutes, and every number arrives with the location it came from so your review is fast and defensible.

Bobyard and a handful of others read site plans too. The difference is the pricing and the depth of the reading. Pilars is $100 per trade, per plan — no per-seat licenses, no annual contract, unlimited projects. And the reading goes past area measurement into the schedules, notes and section details where landscape money actually hides. You can run it on your own set in the demo before you decide anything.

Questions estimators actually ask