Simon 2.0: Proof of Concept Scope

Enabling semantic search on SharePoint using Microsoft 365 Copilot, starting with Jim Mullan's Construction & Development folders.

Prepared: February 25, 2026 Version 1.0 Matt Cooper, Volta Effect
← Back to Projects
1

The Problem

Employees across Southwest Properties struggle to find information stored in SharePoint ("Simon"). The search function returns irrelevant results or misses relevant content entirely. Staff waste time daily searching for documents, asking colleagues, browsing folders manually, or recreating files they can't locate.

"I should be able to say 'show me any correspondence or agreement with Shannix that relates to long term care or assisted living' and then it's going to go through, flag the multiple documents, and give me a summary." Jim Mullan, VP Construction & Development

Today, that search requires downloading files from SharePoint, opening them in Bluebeam, and manually searching within documents across hundreds of pages.

2

Pilot Scope

Department

Construction & Development (Jim Mullan's team)

SharePoint Areas

Jim identified three folder areas for the initial pilot:

  1. Development
  2. Properties and Real Estate
  3. Construction

Users

  • Primary: Jim Mullan
  • Possible second user: One additional team member (TBD by Jim)

In Scope

  • Enable SharePoint indexing for the three designated folder areas
  • Enable Microsoft Copilot search for those indexed areas
  • Enable Copilot search in Outlook for Jim's account
  • Measure search effectiveness over a 3-4 week testing period
  • Document findings and build a repeatable process for other departments

Out of Scope

  • Other departments' SharePoint content (future phase)
  • PII scanning and remediation (parallel cleanup initiative)
  • New portals or tools outside Microsoft's infrastructure
  • Procore integration (Procore is for current projects; SharePoint is historical)
3

Why This Approach

Why Start Narrow

Five discovery calls across four departments revealed varying levels of search satisfaction:

DepartmentSatisfactionNotes
Construction (Jim) Low Searches historical agreements, project files, design documents across 15+ years of content. Can't search within documents.
HR (Jacqueline) Moderate Running a parallel SharePoint cleanup initiative. First department to complete reorganization.
IT (Derrick) N/A (admin) 16 Copilot licenses exist but semantic search features are not configured. Needs support understanding modern SharePoint capabilities.
Finance (Emily) High (~7-8/10) Good naming conventions, well-organized. Search frustrations exist but are manageable.
Accounting (Leslie) High (~7-8/10) Similar to Finance. Strong record-keeping practices minimize the search problem.
Construction has the biggest gap between current experience and what's possible. It's the highest-impact starting point. Finance and Accounting would see incremental improvement; Construction would see a step change.

Why Use Existing Microsoft Tools

Southwest already has 16 Microsoft 365 Copilot licenses. Semantic indexing and Copilot search are capabilities built into these licenses. Rather than introducing new tools or building a separate search portal, this pilot enables features that are already paid for but not yet configured.

  • Avoids creating infrastructure that falls out of sync with Southwest's environment
  • Complements the ongoing SharePoint cleanup initiative rather than competing with it
  • Keeps the technical footprint within what Derrick's team already manages
  • Requires no new vendor relationships or software purchases
4

Implementation

1

Configuration

Week 1 · Owner: Derrick Bertrand, with support from Matt Cooper
#ActionWho
1Enable SharePoint indexing for the three designated folder areasDerrick
2Configure Copilot search access for Jim's accountDerrick
3Enable Copilot search in Outlook for JimDerrick
4Verify Jim can search across indexed content using natural languageMatt + Jim
5Confirm permissions are scoped correctly (Jim sees only what he should)Derrick

Matt is available to support Derrick on any configuration questions or roadblocks.

2

Testing Sprint

Weeks 2-5 · Owner: Jim Mullan

Jim blocks out 30 minutes, 2-3 times per week to test search with real work tasks. During each block:

  1. Search for something specific that comes up in normal work
  2. Note whether Copilot found the right result
  3. If wrong result: note why (wrong version, outdated file, irrelevant)
  4. If nothing found: note what was expected

These notes become the raw input for evaluating the approach before broader rollout.

3

Evaluation

Weeks 5-6 · Owner: Matt Cooper + Jim Mullan

15-minute check-in to review:

  • What worked well
  • What returned wrong or missing results
  • Time savings (qualitative and quantitative if possible)
  • Whether Jim would recommend this for other departments
  • What needs to change before broader rollout

Visual Timeline

Week 1
Config
Weeks 2 - 5
Jim Tests Search
Week 6
Eval
~6 weeks
Total elapsed time
6-9 hrs
Jim's total effort
2-4 hrs
Derrick's configuration
5

Success Criteria

CriteriaHow We Measure It
Jim can find historical documents faster Self-reported comparison: time to find a specific document before vs. after. Target: noticeable improvement in most searches.
Copilot returns relevant results During testing blocks, >70% of searches return useful results in the first attempt.
No wrong content surfaced Copilot does not expose documents outside Jim's authorized access. No PII surfaced from unsanitized areas.
Process is repeatable Configuration steps documented well enough that Derrick can replicate for the next department without Matt's involvement.
Jim recommends continuing At the Phase 3 check-in, Jim sees enough value to advocate for broader rollout.
6

Risks & Mitigations

RiskImpactMitigation
Unclean data returns wrong results Copilot surfaces outdated or duplicate files Scope limited to three folders Jim knows well. He can distinguish correct vs. incorrect results and flag issues for cleanup.
SharePoint cleanup conflict Pilot conflicts with 10-month cleanup initiative Pilot scoped to Jim's areas only, avoiding departments in active cleanup. Jacqueline and Derrick briefed on Feb 18 to ensure alignment.
PII exposure Indexed content may contain sensitive information Indexing respects existing SharePoint permissions. Jim only sees content he already has access to. PII remediation handled by cleanup initiative.
Low adoption Jim doesn't have time to test consistently Testing is lightweight (30 min, 2-3x/week) and tied to real work, not artificial test scenarios.
Copilot limitations Semantic search doesn't work on images, charts, or files over 512MB Set expectations upfront. Construction documents are primarily text-based agreements and correspondence, which are well-supported.
7

Broader Vision

"Really what I'm looking for is something that can search all the different departments. I don't want that to be siloed." Jim Mullan, VP Construction & Development

This pilot is designed as the template for exactly that. Once we prove it works for Construction, the playbook becomes:

  1. Identify the next department's key folders
  2. Derrick enables indexing and Copilot access
  3. A pilot user tests for 3-4 weeks
  4. Evaluate and adjust
  5. Move to the next department
Each department adds to the indexed corpus. This brings Southwest closer to the cross-department search Jim envisions. Departments with strong naming conventions (Finance, Accounting) should be straightforward. Departments still going through cleanup will be sequenced after their reorganization is complete.
8

Stakeholders

JM
Jim Mullan
VP Construction & Development
Primary pilot user. Defines folder scope, provides testing feedback.
DB
Derrick Bertrand
IT Services Manager
Technical configuration, SharePoint administration.
JF
Jacqueline Ferguson
VP People & Culture
Informed stakeholder, alignment with cleanup initiative.
BY
Ben Young
CEO
Economic sponsor, rollout decisions.
MC
Matt Cooper
CEO, Volta Effect
Pilot facilitation, technical support, evaluation.
9

Next Steps

  • Derrick reviews this document and confirms he has what he needs to begin configuration
  • Jim confirms the three folder areas are correct and complete
  • Derrick enables indexing and Copilot search for Jim
  • Jim runs first test searches and confirms it's working
  • Matt schedules the Phase 3 evaluation check-in (target: 4 weeks from go-live)