See the delay
This week, the 8 projects total 41 days past service level. The longest hold, 9 days, is the villas' concept design gate with the design consultant; the project's delay costs about AED 20,700 a day.
Demo · Real estate
Follow one stalled gate to a governed decision and a fix, then one improvement across every project.
An illustrative villa development stalls at a design gate until the Design Director decides. Then test moving one recurring check to an AI agent. This is the programme the interactive demo follows.
Respond to a stalled gate
One stalled gate, from the delay to a governed decision and a fix.
A property developer runs 8 projects through approval gates, each with named roles and agreed service levels. One of them, a villa development valued at AED 142M, is at its concept design gate, which it passes only with a complete package from an outside design consultant.
Illustrative example. Fictional data. Days are calendar days, and today is a Wednesday. Amounts in AED, for illustration only.
-
Portfolio 8 projects Delay this week: 41 days past service level
Villa development · concept design gate
Held by the design consultant, an outside party, for 9 days
Service level for this step: 5 days
Gate 9 days past its due date
Cost of delay AED 20,700 per day (AED 36.5M capital deployed × 7.5% + AED 142M project value × 3.39%) ÷ 365 project figure, counted once; the rates are assumptions
The consultant's hold: 9 days
- 5 days within its service level
- 4 days past it, 3 of them this week
Holding gates past their due date
- Design consultant Outside party · Villa development 9 days
- Design reviewer Another project 6 days
- Commercial reviewer Cost plan, another project 4 days
Delay this week counts the days each open step spent past its service level, Monday to today.
Show the full screen Show less
-
The readiness check
An AI agent checks the consultant's package in 38 seconds and finds 3 issues, by rule and by reading the documents. It recommends a return; this check is shown in the demo.
Readiness check: concept design gate
Gate Readiness Agent · finished in 38 seconds 3 findings
- Planning alignment answered "No", with no comment. Rule; confidence high
- Master plan answered "No", with no modifications listed. Rule; confidence high
- 420 units in the concept design report (p. 14) against 340 in the development brief (p. 3); the report's Assumptions and Deviations section (p. 22) does not declare the extra 80. Reading; confidence medium, 0.82
-
Check and recommendation
Gate Readiness Agent AI agent
3 findings; recommends a return to the design consultant
Show the full screen Show less
-
The authorised person decides
The agent recommended a return; only an authorised person decides. Above AED 50M, as for this AED 142M project, the Design Director approves it, and the audit record keeps who, when and why.
Decision chain: concept design gate
-
Check and recommendation
Gate Readiness Agent AI agent
3 findings; recommends a return to the design consultant
-
Decision
Design Director Authorised person
Approved the return, with the reason recorded
Why the Design Director The project is valued at AED 142M. Above AED 50M, the Design Director decides.
Four kinds of action
- Routine action under an agreed rule The reminder about a cost plan 4 days late
- Check, recommendation or draft The agent's readiness check
- Decision by an authorised person Returning the package to the consultant
- Exception and escalation The escalation to a manager: prepared by the agent, sent once a person confirms it
Show the full screen Show less
Audit record: concept design gate
Entries, in order
- Today, 09:15 Checked the package: 3 findings; recommended a return Gate Readiness Agent
- Today, 10:40 Approved the return, with the reason Design Director
Decision record
- Decided
- Design Director
- When
- Today, 10:40
- Decision
- Return the package to the design consultant
- Seen
- The agent's findings, with the cited pages of the report and the brief
- Reason
- Planning comment and master plan modifications missing; the extra 80 units not declared
-
-
The consultant fixes and resubmits
The consultant opens the return from an email link, with no login or licence, fixes the 3 items and resubmits the same afternoon. The clock stops on them; their page is shown in the demo.
Opened from an email link: no login, no licence
Villa development · concept design gate
Returned by the Design Director: 3 items to fix
-
Finding 1
Add the comment a "No" on planning alignment needs
Comment added
-
Finding 2
List the master plan modifications
Modifications listed
-
Finding 3
Declare the extra 80 units, or revise the layout
Declared on p. 22 of the revised report
Submitted today at 15:05 The clock stops on the consultant
Show the full screen Show less
-
-
Coordinate and measure
Delay this week falls from 41 to 38 days, and reminders go out under agreed rules. In a pilot, each measure has an agreed baseline, target and forecast; actual results come only from real operation.
After the resubmission
What happens next, and who acts
- Today, 15:05 Resubmitted with every item fixed; the clock stopped on them Design consultant
- Today, 15:10 Routine reminder to a Commercial reviewer: a cost plan on another project, 4 days late SLA Watcher, an AI agent
- Tomorrow, 15:10 If it is still late, an escalation to the Commercial reviewer's manager, prepared by the agent and sent only once a person confirms it SLA Watcher, then a person
Reminders go out automatically under the agreed rules; an escalation is sent only once a person confirms it.
Delay this week, 8 projects
- Before
- 41 days past service level
- Now
- 38 days: the consultant's 3 days this week are off the count
- Holding the most
- Design reviewer, 6 days · Commercial reviewer, 4 days
Show the full screen Show less
Improve the operating model
One recurring check, tested across every project.
A separate question: not how to respond to one stalled gate, but which checks recur on every project and who should do them. In the simulation, the baseline comes from 6 months of history across the 8 projects.
Illustrative example. Fictional data; the figures are simulated projections, not results.
-
A check that recurs on every project
Before each project's initiation business plan gate, a Commercial reviewer checks the cost plan by hand, line by line. It is a check, not a decision: the Segment CEO signs off the gate.
Where the work goes: the cost plan's first check
6 months of history, 8 projectsThe programme's recurring work, 100 points The first check: 10 points
Checking each cost plan before its initiation business plan gate, about 40 line items
A check, not a decision: the Segment CEO still signs off every initiation business plan gate.
-
A specific change to roles
In the simulation, a cost-plan agent runs the first check and the Commercial reviewer reviews what it flags, plus a sample. The Segment CEO keeps the sign-off, and nothing changes until they decide.
What-if: the cost plan's first check, every project
6 months of history, 8 projectsChange being tested
-
The cost plan's first check, before each initiation business plan gate
Now Commercial reviewer Person
Moves to Cost-plan agent in Commercial; the Commercial reviewer reviews what it flags AI agent
Stays with a person
Initiation business plan gate sign-off
Segment CEO Authorised person
Applies to every project. Nothing changes in operation until the Segment CEO decides on the evidence.
-
-
Forecast assumptions
The projection states what it rests on: where each figure comes from, what a person still reviews, and what stays the same.
What-if: the cost plan's first check, every project
6 months of history, 8 projectsAssumptions behind the forecast
- Baseline from 6 months of history across 8 projects
- The agent's timing, flag rate and share of issues found come from the simulation's sample
- About 40 line items per cost plan; a tenth of a day for a person to check or review an item
- A person reviews every flagged item and 1 in 8 of the items the agent passes
- The cost plan's contents and the approval rules stay the same
- The Segment CEO keeps every gate sign-off
-
Projected benefits and trade-offs
The projection shows the effect on the first check's days, the issues it finds and AI agents' share of the work. The cost: the reviewer's time on every flag and on the sample.
What-if: the cost plan's first check, every project
6 months of history, 8 projectsProjected impact, per cost plan Measure Baseline Projected First check, cost plan received to checked 4.0 days 1.5 days 2.5 days shorter Issues found at the first check 88% 91% AI agents' share of the programme's work 18% 28% In return: the Commercial reviewer's review of flags and a sample No baseline 1.2 days Show the full screen Show less
-
How actual results will be measured
After the next gates, we measure each result the same way as the baseline and compare it with the target and the forecast. Actual results come only from real operation, so this example shows none.
The tested change, every project
Measure Baseline Target Forecast Actual First check, days per cost plan 4.0 3.0 1.5 After the next gates Issues found at the first check 88% 88% or more 91% After the next gates
Want to see it with your own gates?
In a conversation, we show the interactive demo, which follows this programme, and map your departments and approval rules onto ZedNerve. A pilot then runs on one real process for about 12 weeks, in an environment agreed with you.