Scope Drift Detection
Scope drift detection monitors project emails, meeting minutes, RFI logs, model revisions, and the signed contract scope to catch the design quietly expanding beyond its contracted boundary. It outputs a structured drift report with citations back to each source document and is currently shipping in Beta with 1–3 design partners, including A-2 contracted.
- Early signals of scope drift surfaced before a change order is owed, while the team can still reset expectations with the owner.
- Each drift flag cites the source email, meeting minute, RFI entry, or model revision plus the relevant contract clause.
- Drift reports tie every flagged change back to the original contracted scope, exclusions, and deliverables list.
From “we ate the change” to “we caught the change”.
Workflow today
-
01
Project signed at scope X
Week 0. The team signs a contract with a defined scope of services, a PDF scope exhibit, and an Excel deliverables list. The PM saves the documents in a project folder and walks the team through the highlights in a kickoff meeting. After that, the scope record sits untouched unless someone remembers to reopen the files.
Week 0 -
02
Scope quietly expands
Week 1+. The owner emails “small adjustments” to layouts, facade options, or coordination responsibilities. RFIs start to touch on items that sit outside the original scope, but the team answers them to keep the project moving. Design revisions accumulate in Revit or Rhino without anyone checking against the scope exhibit.
Week 1+ -
03
PM realises late
Month 3+. The project runs hot. Hours burn faster than forecast. The PM realises the team has been doing work that feels out-of-scope but struggles to prove it. Senior staff dig through email threads, RFI logs, and model histories to reconstruct who asked for what and when. The contract scope stays as a static PDF on the server.
Month 3+ -
04
Change order argued retroactively
Month 4+. The firm prepares a change order based on incomplete reconstruction of the project record. The owner disputes which changes were agreed as part of the base scope versus extras. The firm either eats the cost to preserve the relationship or enters a difficult negotiation with limited documentation. Both outcomes burn time, goodwill, and fee.
Month 4+
Workflow with VitruAI
-
01
Set up the agent against the contracted scope
Week 0, ~1 day. The PM uploads the signed agreement, scope exhibits, and deliverables schedules in PDF or Word. The same backbone that powers the Scope Agent and the Contract Agent parses the documents into a structured list of services, phases, and explicit exclusions. The PM can review and adjust the extracted scope baseline before monitoring starts.
Week 0, ~1 day -
02
Agent monitors the project record
Ongoing. The agent connects to a project-specific mailbox or shared folder and reads incoming emails that relate to the job. Meeting minutes arrive either as uploads or via the meeting-minutes-to-action-items workflow if the firm runs it. RFI logs sync from the firm’s chosen tool, and model revisions from Revit or other BIM tools contribute metadata about design changes. Potential drift signals are flagged as they appear, not months later.
Ongoing -
03
PM reviews flagged signals
Weekly. The PM opens a drift dashboard that lists each flagged signal with a short summary, the source document, and the specific clause or exclusion it appears to stretch. Each item shows the original contract language alongside the new request or change. The PM decides whether the item is in-scope, change-order-worthy, or noise, and that decision feeds back into the model for future tuning.
Weekly -
04
Change orders proposed early
When warranted. The PM raises scope issues during regular owner check-ins instead of after the fee is exhausted. Each proposed change order carries citations back to the contract clause, the triggering email or RFI, and the related design revision. When disputes escalate, the same record can support a dispute evidence timeline built with the Memory Agent for claim-ready documentation.
When warranted
Scope drift detection — common questions
-
How does the agent read the contract scope?
At project setup, the PM uploads the signed agreement, scope exhibits, and any schedules of services in PDF or Word. The same contract-reading backbone used in contract clause monitoring extracts structured deliverables, scope boundaries, and exclusions. The PM can edit this extracted baseline so scope drift detection runs against the firm’s accepted interpretation, not an unchecked machine read.
-
Does it read every project email?
Email monitoring is configured per Beta engagement so it only reads the messages that matter for a given project. Typical setups connect to a project-specific mailbox, a shared folder, or labelled threads, and the agent scans for language that suggests new services, extra coordination, or design changes. It stores only the minimal content needed for citation in drift reports, not a full copy of the mailbox.
-
What about meeting minutes and RFIs?
Yes, meeting minutes and RFIs are core inputs for scope drift detection. Minutes can be uploaded directly or passed through the firm’s meeting-minutes-to-action-items workflow, which already structures decisions and asks. RFI logs sync from the firm’s chosen tool so the agent can see when questions touch on out-of-scope responsibilities. Model-revision metadata from Revit, Rhino, or other design tools adds another signal when geometry or scope changes without a matching contract change.
-
How accurate is drift detection?
The agent flags potential drift; the PM still makes the commercial call. During Beta, each design partner tunes thresholds and patterns against their own historical change-order record so false positives drop over time. Every Beta deployment ships a per-project accuracy report calibrated to the customer’s pipeline, and the firm can compare flagged items against actual approved change orders to judge value.
-
How do I join the Beta cohort?
The current cohort is 1–3 design partners, including A-2 contracted, running live projects with structured contracts and a clear scope baseline. Firms need active projects with defined scopes, accessible project email and RFI systems, and a PM willing to review weekly drift reports. You can apply through the “Apply to the Beta” CTA; the engagement runs under MSA + Appendix and can pair the Scope Agent with the Contract Agent for shared monitoring and dispute evidence timelines backed by the Memory Agent.
Want early access?
Pair the Scope Agent with the Contract Agent — both currently shipping with the design-partner cohort.
Apply to the beta