Use four decision criteria
Judge each task on four criteria: input stability, rule clarity, consequence of error, and review effort. A task with stable inputs, clear rules, low consequence, and a short review is a strong candidate for delegation. A task with unstable inputs or a high consequence needs the estimator even when the steps look routine.
Score the criteria before the workload conversation so the decision is not driven by who is available. The cost estimator service page describes the production work that often fits this model, and the bid folder access guide shows how to limit access for the tasks that do move. The criteria are a starting point, not a permanent label, because a task can become delegable after the rules improve.
Score the criteria with a simple high, medium, or low mark and keep the result in the brief. Revisit the score when the project type changes or the review history improves. Input stability is easier to judge with examples than with a score.
A drawing set that arrives complete and dated is stable, while a set that arrives in pieces with verbal changes is not. Rule clarity means a new assistant can apply the rule without asking which interpretation is current. Review effort means a sample of five rows can confirm the output in less time than producing it.
Sort tasks by risk
Low-risk tasks include drawing log upkeep, quantity takeoff from clear drawings, quote collection, and comparison sheet formatting. Medium-risk tasks include scope gap checks and assumption logs, where a missed item can change the bid. High-risk tasks include final markup decisions, bid strategy, and any judgment that depends on the contractor relationship.
Write the risk level beside each task in the brief so the reviewer knows how much sampling the handoff needs. Risk is not the same as difficulty. A complex takeoff can be low risk when the rules are clear and the review is strong, while a simple scope exclusion can be high risk when it changes the bid basis.
Sort by consequence, then decide how much review the task needs. That distinction keeps the team from treating every complex takeoff as a reason to keep the work in house. A medium-risk task can move when the review sample is strong and the brief names the stop conditions.
Start with the least complex package in that category and compare the rework to the in-house baseline. If the rework stays low for two cycles, widen the scope one package at a time. The decision should follow evidence from real bids, not a preference for keeping work close.
Identify handoff-ready work
A task is handoff-ready when the input arrives in a known format, the output has a defined shape, and a reviewer can check it in a short sample. Write the brief and run one pilot before adding the task to the regular queue. The pilot should include a real deadline and a real review, not a training exercise with a cleaned-up drawing set.
A short sample at the end is part of the definition, not an optional extra. If the pilot needs constant rescue, the task is not ready. Either the rule is missing or the input is too unstable.
Fix the brief or keep the task in house for another cycle. A failed pilot is useful evidence, not a reason to push harder.
Name what stays in house
Keep final scope interpretation, bid strategy, owner negotiation, and any decision that changes the commercial position with the estimator. These tasks depend on context that is hard to write down and expensive to get wrong. The assistant can prepare the evidence for the decision without making it.
Write those tasks on the same list as the delegated work so the boundary is visible in one place. Also keep the first review of a new project type in house until the rules are proven. Once the team has a brief and a review history, parts of the work can move.
The AACE recommended practices frames clear roles and records as basic management discipline, and that discipline matters more when the work crosses a company boundary. That first review is where the estimator learns whether the brief is strong enough to hand off.
Run a small pilot
Pick one task and one deadline for the pilot. Define the output, the review sample, and the success measure before the work starts. At the end, compare the rework and the estimator time saved against the baseline.
A pilot that saves time but creates rework is not a success yet. Decide after the pilot whether to expand, adjust, or stop. Write the decision down with the evidence so the next conversation starts from facts.
The goal is a stable split of work, not a permanent ban on either side. A stopped pilot is cheaper than a permanent task that never quite works. Choose a pilot task that already has a written brief and a reviewer who can answer questions during the week.
Measure the estimator time before and after the pilot, not only the assistant output. Include the review time in the comparison so the saving is honest. A pilot that shifts work without reducing total effort is still useful evidence, but it is not a capacity gain.
| Task | Risk | Decision |
|---|---|---|
| Drawing log upkeep | Low | Delegate with a receipt rule |
| Quantity takeoff | Low to medium | Delegate with a sample review |
| Scope gap review | Medium | Prepare only, estimator decides |
| Final markup | High | Keep in house |
