AGC NYS 2026 AI sessions: questions for a controlled preconstruction pilot
Build a controlled AI preconstruction pilot around AGC NYS 2026 with permitted files, a named reviewer, a retained error log and clear criteria for stopping.
Author
Ido Gedanken, CEOPublished

Bring a current or completed project and see how Piper saves review time, surfaces scope gaps, and applies your company's knowledge.
Book a demo- Type
- Guide
- Read
- 2 min
- Published
On this page
Before adopting an AI workflow discussed at AGC NYS, a chief estimator should name the output, the human reviewer and the conditions that stop the pilot. Test a small, permitted package against known evidence, and retain the errors as well as the successful findings.
This independent Piper guide was checked November 15, 2026. AGC NYS lists its conference December 8–10, 2026 in Saratoga Springs. Its program advertises an AI keynote and sessions on business use and New York construction AI compliance. Consult the organizer for registration and the current schedule. Session topics do not establish legal requirements or future takeaways.
Write the pilot charter before selecting a result
Illustrative task: compare a scope against a later addendum affecting a mechanical penetration. The desired output is a list of possible scope changes and unresolved trade ownership, with references for a reviewer. It is not authority to issue a scope or alter a bid without review.
| Charter field | Decision to record |
|---|---|
| Permitted files | Specific approved versions, redactions and who authorized their use |
| Output and boundary | Addendum-to-scope findings for one package, not an autonomous bid |
| Known limitations | Missing drawings, incomplete quote coverage and any untested file types |
| Reviewers | Estimator checks sources; chief estimator owns disposition |
| Escalation | Route ambiguous trade ownership to the responsible bid lead |
| Error log | Record missed changes, unsupported findings, wrong versions and corrections |
| Retained evidence | Keep inputs, version identifiers, output, reviewer decision and source references under company policy |
| Stop criteria | Stop for unauthorized data, an untraceable material finding or inability to review before the decision deadline |
Before the test, identify at least one known change and one genuinely unresolved issue. After the output, the reviewer checks the actual addendum and original scope, records whether the item was found and corrects the disposition. Separate a missed finding from a source that was not present in the permitted input. A pilot result is evidence about that defined package, not an industry benchmark.
Bring compliance questions to the right people
Ask what company approvals govern the chosen files, what retention and access terms apply, and who should assess applicable legal obligations. Take these questions to company counsel and the responsible information-security staff; do not infer a rule from a workshop title. Resolve permissions before placing project or subcontractor information into a demonstration.
Use the AGC event plan for team allocation and the New York Build demo test for a distinct vendor demonstration. The full AI software evaluation guide covers the broader purchasing decision.
Write reusable review questions in the form described by the pre-bid questionnaire guide. Separate questions the documents can answer from permissions, approvals and judgments that still belong to people.
Keep company questions attached to project evidence
Piper Survey applies company questions to project documents, answers supported questions and flags items needing human response. A pilot owner can evaluate that workflow against the charter above; the accountable reviewer still decides the project's treatment.
Related reading
How to evaluate AI preconstruction software
Test AI preconstruction software on representative documents, company standards and revisions. Inspect findings, source accuracy and the effect of an addendum on prior work.

Surveys: the questions worth asking before you bid
A standing pre-bid questionnaire turns the things your best estimator always remembers to check into something the whole company checks every time. How to build one, who answers what, and how to keep it alive.

Piper removes manual review from the critical path and brings project data, company knowledge, and expert checks into every preconstruction decision and workflow
See Piper on your project
Bring a current or completed project and see how Piper saves review time, surfaces scope gaps, and applies your company's knowledge.