Service tasks and execution checklists

Every service in your catalog can carry two parallel lists — materials your crew consumes and tasks your crew performs. A Kerdi shower install service knows it needs membrane, thinset, and screws (materials) AND that the crew will demo the old surround, frame the curb, set the pan, install the membrane, then set tile (tasks). Both lists travel with the service. When a client signs a proposal that uses the service, the tasks land on the project automatically — already broken out, ready to assign.

You author tasks the same way you author materials: open any service in your catalog and scroll to the Tasks section below Materials.

How a task is shaped

Each task has a title, a typical-hours estimate, and an optional scaling rule. Reading top to bottom, a row reads like a sentence:

Set the tile field · Takes about 1 hr · scales by area · at 10 sqft per hour

That sentence means: this task takes about an hour of setup time, plus the crew can set roughly 10 sqft of tile per hour after that. On a 40 sqft floor, the task lands as ~5 hours. On a 100 sqft floor, ~11 hours. The math runs automatically when the proposal is signed — you don't have to recalculate per project.

The dropdown has four options:

• doesn't scale — flat hours regardless of project size. Right for site setup, walkthroughs, final clean, meetings, and per-item work like framing a single niche (just put the per-item hours in typical hours; most projects have 0-2 of these things anyway).
• scales by area — sqft per hour. Right for demo, set, grout, seal. The specific surface (floor vs. wall) is figured out from the parent service's pricing setup.
• scales by length — linear feet per hour. Right for edge trim, baseboard, caulk lines, curb work.
• a custom formula… — escape hatch for the genuinely weird ones. You can reference any project dimension directly.

The production-rate framing ("20 sqft per hour", not "0.05 hours per sqft") is on purpose — it matches how contractors actually talk about their work. Higher number = faster.

Skills and SOPs travel with the task

Each task row can carry:

• Required skills — purple chips. Picked from your team's skill vocabulary. When you assign tasks to crew, BidCraft suggests the right people first.
• Attached SOPs — green chips. The crew opens the task on a job and can tap to read the SOP without leaving the app.

Both are optional. Tasks without skills or SOPs work fine; the chips just narrow who can do the task and what they should read first.

Chain dependent tasks

Each task row in a service template has an After (in this service) multi-picker. Use it to declare which other tasks in this same service must finish before this one can start. Common examples: tile install depends on waterproofing; grout depends on tile install.

Dependencies authored at the service-template level are inherited automatically on every project that uses the service — set them once, every future job benefits. Project-level overrides are also possible from the task Details on a specific project.

The schedule on a project page topologically sorts by these dependencies, so changing one task's date reflows the dependents automatically.

Generate with AI

If a service is mostly empty — name and price are set but no tasks yet — hit Generate with AI at the top of the Tasks section. Biddy proposes 3-7 tasks with realistic typical hours and scaling rates calibrated to a typical 2-person crew.

[!note] AI-picked rows show a small ✨ badge reading "AI pick — verify or swap" (it also names whether the pick came from a suggest run or a proposal build). That's your cue to either confirm Biddy nailed it (one-tap Confirm) or edit the row. A manual edit counts as verification.

There's a Verify all action at the top of the Tasks list when there's a batch of AI picks waiting. Use it when the AI got close enough that you just want to bless the whole list at once.

Preview before you save

Open the Show formula test panel toggle at the bottom of the Tasks section. Type in sample dimensions (a typical floor sqft, perimeter, etc.) and every row's preview updates with what the crew would actually see on a project of that size. Same math the AI runs at sign time — what you see in the editor is what'll land on the task.

Next steps

• See how tasks land on a project after a client signs: Tasks on signed projects
• Build org-wide checklists that fire on every signed project (Day 1 walkthrough, closeout punch list): Org-wide checklists
• Set up the materials side first: Service materials and favorite-for-roles