top of page
Collier & Associates brand mark featuring the core services: Business Transformation; Enterprise Systems; AI-Enabled Execution.
Company Type

Collier & Associates practice-built, engagement-linked; built and operated within the practice and applied to engagement delivery.

What Drove the Need

  • Enterprise-systems assessments, business cases, and vendor selections generate more evidence than a principal can process by hand at deal speed

Business Impact

  • Four ERP vendors scored on 1,012 Must-Have points from 880+ weighted requirements — IFS at 95% with zero customization — with undisclosed third-party dependencies surfaced in three vendor responses (Case 54)

  • 722 teammate demo observations sentiment-scored per vendor and demo section to complement RFI scoring in a time, equipment, HRIS, and payroll selection (Case 55)

Systems/Data Integration

  • Applied to multiple enterprise-systems assessments, business cases, and value-creation roadmaps (2024–present); RFI framework applied across five selection engagements

Design

  • 47a — transcript and notes processors with transformer sentiment scoring; e.g., 722 teammate observations across vendor × demo section scored and summarized for a construction-services selection (January 2025 demos)

  • 47b — stepped, template-driven evaluation system: three-level APQC-aligned process taxonomy (~240 activities) that every requirement keys to; ~735 functional, ~65 non-functional, ~83 technical requirements; weighted criteria (Functionality 30%, Technical Compatibility 25%, Implementation Timeline 20%, Cost 15%, References 10%); vendor scoring with Must-Have coverage, coverage gap, and customization-to-score metrics; scripted demo scoring; sentiment analysis of demo comments

  • 47c — brand-styled document generation, PDF text extraction, Zoom/Teams transcript processing against a corpus, QuickBooks data analysis, Excel financial-model generation

Technology Stack

  • Python 3.12; LangChain/LangGraph; RAG (FAISS/ChromaDB); HuggingFace Transformers and BART zero-shot classification; spaCy; scikit-learn

  • OpenAI, Anthropic Claude, and Google Gemini APIs; Claude Code

  • FastAPI, PostgreSQL, DuckDB; pptxgenjs / docx / openpyxl generation pipelines

Applied AI Application Set

Three sets: (a) IT due-diligence notes, meeting/audio transcript, and sentiment-analysis processors; (b) the PARA™ Vendor/System RFI Framework with embedding-based semantic requirement matching and human checkpoints; (c) corpus-driven PowerPoint, Word, and Excel generation pipelines.

bottom of page