From the College Athletics edition of October 8, 2026
Mississippi State Athletics Communications scaled an AI-assisted stats pipeline from soccer to multi-sport, cutting turnaround from hours to minutes
Mississippi State University Athletics / MSU AI Hub · NCAA Division I athletics communications and statistics operations (SEC) · United States (Starkville, Mississippi)
- Publisher
- MSU AI Hub
- Original publication
- October 6, 2026
- Source retrieved
- 2026-10-09
What happened
MSU Athletics Communications used ChatGPT to draft Excel and Power Query formulas that pull and clean NCAA game data, then hit rate limits that risked account lockouts and still left full-player refreshes near an hour. Working with doctoral candidate Josh King in the athletics AI & Data Science Lab, staff routed underlying data files straight into spreadsheets, cutting refreshes so a full slate of player statistics turns around in minutes instead of hours. Will Simmons removed a redundant career-totals refresh by deriving career numbers from game logs. The system began with soccer, moved to basketball, with softball next, and relies on sport-specific scoring quirks (for example baseball innings as thirds). Ogden stresses staff are not coders and credits collaboration; next focus is teaching student workers to build similar fixes. No controlled baseline study, error rates or cross-school replication is reported.
Why it matters
Why now: published Oct 6, 2026 and not in the archive before this edition; it is an operational, in-house AI automation story for athletics communications rather than another FanWord-style vendor announcement. Interpretation: departments can gain real turnaround wins by pairing commodity LLM help with campus data-science capacity and existing NCAA XML feeds, without buying a content-generation platform.
Evidence and measured results
University AI Hub article dated Oct 6, 2026 quoting Brian Ogden (director of communications) and describing collaboration with Josh King and Will Simmons. Self-reported efficiency (minutes vs hours) and process history; no independent measurement, sample sizes, accuracy audit or cost figures.
Limitations and uncertainty
Institutional case study; efficiency is self-reported; no accuracy metrics; ChatGPT role limited to formula drafting; not an independent evaluation.
Put this evidence to work
Lighthouse Advisory interpretation, grounded in this source. Enriched 2026-10-08; this does not change the original publication date. Each role follows the same checklist so takeaways compare across sources.
Sales
Role takeaway
- Customer problem
- Athletics communications staff spend hours each game night refreshing player and career statistics across sports.
- Stakeholders
- Director of athletics communications
- Athletics data/AI fellow or campus data science
- Sports information directors
- Student communications workers
- Athletics IT
- Discovery questions
- How long does a full player-stats refresh take today?
- Do you have direct access to NCAA or provider data files?
- Which sports share the same underlying file format?
- Who maintains formulas when student workers turn over?
- Value hypothesis
- Routing existing competition data files into a shared spreadsheet pipeline may cut refresh time from hours to minutes when formulas encode sport rules correctly.
- Potential engagement
- A two-sport stats-pipeline readiness assessment with timed baseline refresh and box-score reconciliation checks.
- Unsupported claims
- Vendor content-AI tools produce the same efficiency
- Measured accuracy across all sports
- Cost savings figures
- Replication outside MSU
Pre-sales engineering
Role takeaway
- Fit
- Departments with access to NCAA/provider XML or data files and staff who can maintain Excel/Power Query or equivalent transforms.
- Architecture and integration
- LLM-assisted authoring of spreadsheet transforms; direct file ingest into athletics stats workbooks; shared formulas adapted per sport.
- Prerequisites
- Access to underlying data files
- Named technical co-owner
- Sample historical files for testing
- Constraints
- Rate limits if scraping websites
- Sport-specific scoring quirks
- Staff are not professional software engineers
- Security
- Use least-privilege data accounts; avoid high-frequency public scraping; keep credentials out of shared student workbooks where possible.
- Proof of value
- Timed full-roster refresh before/after
- Box-score totals match official stats on a sample
- Career totals reconcile to game logs
Delivery
Role takeaway
- Work
- Operate and extend a multi-sport stats refresh pipeline with documentation and student-worker training.
- Dependencies
- Athletics communications
- Campus AI/data science capacity
- Stable data-file access
- Ownership
- Director of athletics communications with the athletics data/AI fellow.
- Skills and adoption
- Teach student workers to adapt formulas for new sports rather than only run refreshes.
- Governance checkpoints
- Sample-data test before each new sport
- Version-controlled formula change log
- Game-night reconciliation checklist
- Acceptance criteria
- Proposed: full-slate refresh completes in under 15 minutes for live sports
- Proposed: 100% of sampled player totals match official box scores before publish
- Risks
- Silent formula errors across sports
- Knowledge loss when student workers leave
- Account lockouts if scraping returns
Implementation considerations
Lighthouse Advisory interpretation across the operating dimensions a public-sector buyer must settle before this evidence becomes a design. Each note answers the question under its heading for this specific source.
Architecture and integration
What must connect, and where does the AI sit in the workflow?
LLM-assisted formula authoring over Excel/Power Query, then direct ingest of athletics XML/data files into spreadsheets instead of scraping the NCAA site. Interpretation: the durable architecture is the data-file route and shared formulas across sports, not the chatbot used to bootstrap formulas.
Governance
Who approves, reviews and stays accountable for outcomes?
In-house scripts touching NCAA competition data need version control, access controls and a named owner so student worker turnover does not orphan the pipeline.
Security and privacy
What data, permissions and controls need testing?
Stats pipelines use public competition data; still require controlled accounts so automated refresh cannot lock credentials or pull beyond authorized scopes.
Accessibility and workforce
Who is affected, and what skills or accommodations follow?
Shifts student-worker learning toward building and adapting automation, not only manual stats entry.
Procurement
What should contracts, pricing and exit terms secure?
May reduce need for paid stats-refresh tooling if file access and campus data-science help are available; still budget for maintenance time.
Operating model
Which teams own the service once it runs?
Communications and an athletics AI lab co-own an operational pipeline, blurring SID and data-engineering roles.
Publication history
- 2026-10-08College Athletics · Issue 324 resources
Stable resource ID: msu-athletics-ai-stats-system-every-sport-2026-10-06