Schedules & BOQ Text

A schedule is a table that says the same kinds of thing about many similar items. A door schedule has a row per door and columns for reference, location, size, material, core, finish, fire and acoustic rating, glazing, ironmongery set and remarks. A window schedule does the same for openings; a finishes schedule does it for floors, skirtings, walls and ceilings room by room. The value of a schedule is entirely in its discipline: every row complete, every column used the same way, the same term for the same thing every time, and every reference matching the drawings.

That discipline is exactly what a tired human loses at row forty and what Claude never loses. Give it your columns and a handful of example rows and it will extend the pattern consistently, keep your naming conventions, flag rows where a field is missing, and spot where you have called the same finish two different names. Ask it to "check this door schedule for consistency: list any door where a field is blank, any finish named more than one way, and any fire rating that looks inconsistent with the door's location" and it becomes a genuinely useful second pair of eyes on your own draft.

A good prompt gives Claude the structure and lets it populate the wording, not the facts:

Here are my door schedule columns and three example rows. Extend the table for
the remaining doors I describe below. Keep my exact naming for materials and
finishes. Where I have not given you a value, write TBC - do not guess a size,
rating or ironmongery set. Flag any row that looks internally inconsistent.

Work in an Artifact so the table stays editable beside the chat, and keep a master column definition in a Project so every schedule across the job uses the same headings and conventions. The result is schedules that are more complete and more consistent than hand-built ones - with the facts still coming from you.

A bill of quantities item has two parts: a description and a quantity (with a unit and, later, a rate). The description is careful, standardised prose - "Providing and laying 20mm thick cement plaster in CM 1:4 to internal walls, including scaffolding, curing and finishing smooth" - and it usually follows a standard method of measurement so that everyone pricing the job reads items the same way. Writing these descriptions well, consistently, across hundreds of items is slow, and it is precisely where Claude helps.

Give Claude your specification and your item list and it will draft clean, consistent descriptive text, keep the phrasing parallel across similar items, and include the "including..." clauses that make an item complete and unambiguous. Ask it to align the descriptions to your method of measurement conventions and to keep units consistent. This can turn a day of item-writing into an afternoon of drafting and checking.

But the quantity is another matter entirely, and here the rule is absolute: Claude does not measure your building, and any number it offers as a quantity is invented. Quantities are taken off the drawings by measurement - by you, your team, or a quantity surveyor - according to a defined method. They are the priced core of the document; a wrong quantity is a wrong price and, in many contracts, a real financial exposure. Never ask Claude to "estimate the quantities" and paste the result into a BOQ. It has not seen your drawings to scale, cannot measure, and its arithmetic can be wrong even when the method is right.

What Claude can do around quantities is set up the calculation, not certify it: help you structure a take-off sheet, write the formula for a repeating volume, or sanity-check whether a total looks the right order of magnitude given your own inputs. That is Module 7 and Module 8 territory - Claude as a calculator you check, never as the QS who signs the bill. Keep the wording with Claude and the numbers with the people and methods accountable for them.

The subtle failure in documentation is not a single wrong clause - it is drift. The spec calls a product one thing, the finishes schedule another, the BOQ item a third; a door is 900mm on the plan, 1000mm in the schedule; a finish appears in the schedule but has no BOQ line. Contractors price the gaps, and disputes grow in them. Cross-document consistency is tedious to police by hand, and it is one of the most valuable things Claude can help with - if you give it the documents to compare.

You can paste your specification and your schedule and ask Claude to reconcile them: "list every finish named in this schedule that does not appear in this specification, and every spec product not used in the schedule, and any place the two use different terms for what looks like the same material". Because Claude holds hundreds of pages in its context window at once, it can hold the whole set together in a way a human flicking between files cannot. It will surface candidate mismatches fast.

Two honest limits. First, Claude finds candidate inconsistencies; you confirm which are real, because sometimes two different terms genuinely mean two different things and Claude cannot always tell. It is a flag-raiser, not an adjudicator. Second, it only reconciles what you give it - if the current drawing revision is not in the context, it cannot catch a mismatch against it. Feed it the live versions and treat its output as a to-check list, not a clean bill of health.

Used this way, Claude closes the gaps that quietly cost money: the orphaned finish, the size that disagrees with itself, the term that drifted. You still make the call on every flag, but you are making calls on a list Claude assembled in seconds rather than hunting for problems you may never have found. Consistency is where documentation is won or lost, and it is one of Claude's most undersold strengths.

Run this loop and the division of labour stays clean. One: define the structure once. Put your schedule columns, naming conventions and BOQ description style into a Project so every table matches. Two: draft descriptions and rows. Give Claude your specification, your item list and example rows, and let it populate the descriptive text and extend the schedules - with a standing rule to write TBC, never a guessed value. Three: keep quantities out. Do the take-off yourself or with your QS, from the drawings, by your method of measurement, and paste the measured quantities into the document; do not let Claude supply them. Four: reconcile. Have Claude cross-check the spec, schedules and BOQ against each other and produce a mismatch list. Five: verify and own. Confirm each flag, correct the drift, and read the priced document as the professional accountable for it.

The interiors version of this is FF&E and finishes-heavy, and the principle is identical: Claude drafts the schedule wording and the specification-linked descriptions; you confirm every product code, dimension, fabric and finish against the current data sheet, and you or your supplier certify quantities and lead times. A schedule that reads perfectly can still order the wrong quantity of the wrong laminate - fluency is not accuracy.

The honest summary is that Claude gives you back most of the grind of schedule and BOQ text and, as a bonus, a more consistent document than you would build by hand, because it never tires of the pattern. What it must never give you is a quantity. A bill of quantities is a measured, priced, accountable document; the measurement is a professional act with a name attached to it. Let Claude write the sentence that describes the item, and keep the number that prices it exactly where it belongs - with the people and methods that can stand behind it.