AI Team One Executive Decision Samples

SAMPLE REPORT — Illustrative company and figures. No live research was performed, actual agent executions: 0, MAX Not run, Guardians not run (built · integration pending). Human decisions remain required.

Sample Report 6 · Construction

Construction Company — Why Are Projects Over Budget?

Ridgeview Builders (sample) · Sample results only

Client assignment

Five active projects are running over budget. The owner wants the causes ranked, owners assigned and margin recovery options.

How the client assigns it to TED: The owner shares sample budget-to-actual reports for five projects with TED and asks where the money went. TED confirms the analysis will separate recoverable from sunk cost.

Executive summary

The five projects are $970K (7.2%) over a $13.5M combined budget. Materials and labor explain 60% of the overrun; $220K of change orders were performed but not yet billed and are the fastest recovery opportunity.

$13.5M

Combined budget

Five projects

$970K

Total overrun

7.2% over budget

$220K

Unbilled change orders

Fastest recovery

Key findings

  1. 1. Project 3 is the worst by percentage

    15% over on a $1.2M budget.

  2. 2. Materials price drift

    $310K, mostly from purchase orders placed late at higher prices.

  3. 3. Change orders not billed

    $220K of approved field changes lack signed client paperwork.

Budget to actual by project (illustrative)

Variance = actual − budget.

  • ProjectP1

    Budget
    $2.00M
    Actual
    $2.18M
    Variance
    $0.18M
    %
    9%
  • ProjectP2

    Budget
    $3.50M
    Actual
    $3.71M
    Variance
    $0.21M
    %
    6%
  • ProjectP3

    Budget
    $1.20M
    Actual
    $1.38M
    Variance
    $0.18M
    %
    15%
  • ProjectP4

    Budget
    $4.00M
    Actual
    $4.12M
    Variance
    $0.12M
    %
    3%
  • ProjectP5

    Budget
    $2.80M
    Actual
    $3.08M
    Variance
    $0.28M
    %
    10%
  • ProjectTotal

    Budget
    $13.50M
    Actual
    $14.47M
    Variance
    $0.97M
    %
    7.2%

Overrun by cause and mitigation owner

Shares of the $970K total.

  • CauseMaterials

    Amount
    $310K
    Share
    32%
    Owner
    Procurement lead
  • CauseLabor

    Amount
    $270K
    Share
    28%
    Owner
    Superintendents
  • CauseChange orders (unbilled)

    Amount
    $220K
    Share
    23%
    Owner
    Project managers
  • CauseSchedule / general conditions

    Amount
    $110K
    Share
    11%
    Owner
    Operations director
  • CauseProcurement expediting

    Amount
    $60K
    Share
    6%
    Owner
    Procurement lead

Options & decision table

Margin recovery options

Illustrative.

  • OptionDocument and bill change orders

    Potential recovery
    Up to $220K
    Timing
    30 days
  • OptionLock remaining material buys early

    Potential recovery
    Limits further drift
    Timing
    Immediate
  • OptionRe-sequence crews on P3 and P5

    Potential recovery
    Reduces labor overrun
    Timing
    60 days

Prioritized action plan

  • 1Assemble signed backup for every unbilled change orderProject managers · 2 weeks
  • 2Buy remaining materials at locked prices where cash allowsProcurement lead · 30 days
  • 3Weekly labor productivity review on P3 and P5Superintendents · Ongoing

Success metrics

  • Change orders billed within 14 days
  • No project overrun above 8% at completion
  • Weekly cost-to-complete forecasts

SOCRATEE's Perspective

Sample text, not generated

Assumptions to test

  • Clients will accept change-order billing
  • Budgets were realistic at bid

What deserves questioning

  • Is the estimate, not execution, the real problem?
  • How much overrun is still ahead?

Sources & evidence

Illustrative assumptions (not evidence)

  • All project figures are sample data

Verified data needed before deciding

  • Job cost ledgers
  • Signed change orders
  • Supplier quotes and contracts

Verified evidence attached: none. No citations are provided because no research was performed.

Safeguards

  • Contract and billing decisions reviewed by management and counsel where needed

Report status · Research & Review Record

  • Actual agent executions: 0
  • MAX independent grade: Not run
  • 18 Guardians: Built · Integration Pending; no review occurred
  • Live research: None performed
  • Final decision: Always a human decision

Specialist roles (role definitions · not executed)

  • TED — Coordinator: would scope the assignment, route work and assemble the report.
  • RES — Would collect cost records and supplier pricing.
  • KNOLLY — Would reconcile ledgers to project budgets.
  • Project Controls Analyst (role definition) — Would build variance and cost-to-complete analysis.
  • SOCRATEE — Would question the original estimates.
  • LIBBY — Would check contract terms before billing.
  • MAX — Potential independent grader. Not run; no grade exists.
  • 18 Guardians — Potential oversight layer. Built · Integration Pending; no review occurred.

How AI Team One Works — Step by Step (prewritten)

  1. Stage 1

    The assignment

    The owner asks why five projects are over budget.

  2. Stage 2

    TED coordinates

    TED routes variance analysis by project and by cause.

  3. Stage 3

    Specialist responsibilities

    A Project Controls role builds variance, KNOLLY reconciles ledgers, LIBBY checks contract terms.

  4. Stage 4

    Oversight design

    Oversight here would center on whether estimating or execution is at fault: SOCRATEE would test whether the bid estimate, not jobsite execution, is the real driver, LIBBY would route contract and billing questions to management and counsel rather than deciding them, MAX could grade the variance roll-up, and the Guardians could review the owner assignments. In this sample none of that ran: MAX Not run, Guardians not run (built · integration pending).

  5. Stage 5

    The executive report

    The result would look like "Construction Company — Why Are Projects Over Budget?": findings, options, a prioritized action plan and the data still needed. Prewritten sample content; actual agent executions: 0.