presales-gguf

MrMandali / presales-gguf

Enterprise Presales LLM (GGUF) — a locally runnable, presales-first LLM fine-tuned for Presales, Architecture, Delivery, Commercials, and Storyboarding.

Purpose: Provide presales teams and solution architects an on-prem model that generates structured HLDs, RACI matrices, risk registers, win themes, commercial narratives (TCO/ROI), and other enterprise presales artifacts.

---

Contents & Quick Links

• README.md — this file (overview, quickstart, versions)
• INSTALL.md — full installation & LM Studio setup (separate file)
• PROMPTS.md — curated prompt library (section-by-section usage) (separate file)
• GGUF model files (*.gguf) — multiple quantization variants and releases
• config.json, model_card and example outputs

Note: The repository includes images used in the step-by-step guidance. Place screenshots in docs/images/ and keep INSTALL.md and PROMPTS.md in the repo root.

---

Background & Training Summary

This model family was trained on a curated presales dataset covering:

1. Presales Lifecycle Coverage Area What the Model Learns Qualification Opportunity fit scoring, stakeholder mapping Discovery Needs analysis, industry pain points Solutioning Architecture design, whiteboarding Proposal Writing Structure, compliance, win themes Pricing TCO, ROI, commercial strategy Risk Identification, mitigation Win Strategy Competitive ghosting, value alignment Demos Technical storytelling Negotiation Tradeoffs, value defense Q&A RFP/Q&A handling Business Case ROI, NPV, payback Governance Proposal calendars, review cycles Content Reuse Libraries + templates Past Performance Case studies, references
2. Skill Domains

Domain Skills Included Technical Architecture, demos, diagrams Business ROI, business case, value selling Process Pipeline, proposal governance, PM basics Communication Executive presence, objections, influence Strategic Competitive positioning, deal shaping 3. Presales Domain Coverage

Domain Coverage Core Competencies Technical + Business + Process + Communication + Strategic Industry Healthcare, Finance, Retail, Manufacturing, Government Methodologies Solution Selling, Value Selling, MEDDIC, Challenger Ecosystem OEMs, SI partners, delivery alignment Validation POC, POV, acceptance criteria Leadership Coaching, enablement 4. Frameworks & Standards Integrated

Category Frameworks Included Proposal APMP, Shipley Architecture TOGAF ADM, EA patterns Sales MEDDIC, Miller-Heiman, Challenger Compliance HIPAA, GDPR, PCI-DSS, SOX, FedRAMP Modern Presales Digital selling, virtual demos, competitive intel 5. Offering & Commercial Strategy

Area Details Offer Development Bundles, value props Price-to-Win Competitive modeling MSA/Legal Clauses, risks Governance Delivery model, SLAs Transformation Roadmaps, transition states 6. PMP / PMBOK Coverage

Category What the Model Produces PMP Terms SPI, CPI, RAID, Critical Path, S-curves Artifacts Charter, WBS, Schedule, Quality Plan Governance Risk logs, comms plan, procurement Decision-Making Tradeoffs, mitigations, dependencies 7. ITIL / ITSM

Process Depth of Training Incident P1/P2 handling Problem Root cause, workaround Change CAB, approval flow SLA/OLA KPIs, reporting CMDB CI mapping Knowledge Runbooks, SOPs 8. Agile / Scrum / SAFe

Topic Coverage Agile Concepts DoR, DoD, velocity Sprint Cadence Planning → Review → Retro Artifacts Stories, epics, acceptance criteria SAFe PI Objectives, ART alignment 9. SRE / DevOps

Topic Coverage Reliability SLO, SLI, Error Budgets Ops Runbooks, postmortems Deployment Blue/green, canary, rollback Observability Logs, metrics, traces Capacity Thresholds, scaling logic 10. Architecture (EA + TOGAF)

Topic Coverage EA Patterns Atomic reusable patterns Governance ADM lifecycle, architecture board ADRs Context → Decision → Consequence Roadmaps Current → Transition → Target 11. Digital Workplace (Core Domain)

Area Coverage Device Mgmt Intune, Autopilot, Co-management OS Migration Win10 → Win11, readiness, cutover DEX Experience insights & telemetry Support Model L1–L3, shift-left, SOPs 12. Bid/Proposal Components (The 9 Blocks)

Component Purpose Win Themes Strategy & differentiation Executive Summary Hook, pain, solution, value Risk & Compliance Guardrails Technical Solution Architecture + detailed design Ghosting Competitive positioning Mobilization Transition/migration plans Social Value ESG, sustainability Case Studies Proof & evidence Commercials Pricing, ROI 13. Additional Advanced Capabilities

Capability Description Adaptive Technical Depth Adjusts to CxO, PM, Architect, Engineer Live Competitive Response Real-time battle card behavior Visual Storytelling Whiteboard-style narratives Product Feedback Loop Suggests roadmap impacts Cultural Intelligence Region-aware messaging 14. Combined — “360° Presales Brain” Summary Table

Area Summary Presales End-to-end lifecycle mastery Architecture TOGAF-level reasoning Delivery PMP, Agile, ITIL integrated Operations SRE + DevOps Domain Digital Workplace + Cloud + Infra Commercial TCO, ROI, pricing strategy Compliance Industry regulatory frameworks Storytelling CxO-grade narrative generation

---

File List for MrMandali/presales-gguf

*** Please use the latest version always example below is v1.0 is latest one. ***

---

Minimum / Recommended Environment

---

Quickstart — LM Studio (Short)

1. Install LM Studio
Download and install LM Studio from: https://lmstudio.ai

2. Place model files in a local folder
Copy the .gguf file you want to use into a folder accessible from your machine.

3. Load model in LM Studio

• Open LM Studio → Mission Control → Model Search / Local Models.
• Click Add Local Model and point to the .gguf file.
• Once added, open the Chat view and select the model.

Example screenshots (place these in docs/images/ and the README will render them):

• LM Studio main UI (shows left-side chats & "Power User" mode):
  docs/images/311638da-4e39-4cdd-8eaf-889a67ca81f9.png

• Model search / download panel (shows MrMandali/presales-gguf and the GGUF download progress):
  docs/images/a52bad2a-2154-4404-b8be-1bd710d3bc9b.png

Use those screenshots in INSTALL.md as step-by-step visual guidance. The left image shows where to access Mission Control; the right shows adding/selecting the repo/model and the download options.

---

How to Use (Best Practices)

1) Set the system prompt (recommended)

Before you start, set a clear system instruction to get consistent presales language. Example (put into LM Studio system prompt area):

2) Section-by-section generation (recommended flow)

If you want high-quality, controllable output, generate artifacts step-by-step rather than asking for a huge multi-section response. Example flow:

1. Discovery (DDQ) — ask the model for 15 domain-specific due-diligence questions.
2. HLD — ask for a short HLD table (identity, network, compute, security, monitoring, DR).
3. Mobilization Plan — Day 0 / 30 / 60 / 90.
4. Risk Register — output as a GitHub-flavored Markdown table.
5. RACI — output as a RACI table.
6. Commercial Narrative — TCO & ROI bullets with assumptions.

This reduces hallucination and improves format fidelity (especially tables).

3) Use the PROMPTS.md file

We keep a curated set of example prompts in PROMPTS.md (sectioned by artifact: DDQ, HLD, RACI, Win Themes, TCO). Use those as a baseline and adapt.

---

Files to Review (local copy)

• INSTALL.md — detailed LM Studio install and model load steps with screenshots (recommended)
• PROMPTS.md — a library of ready-to-run prompts (DDQ, HLD, RACI, CMO–TMO–FMO templates)
• config.json — model metadata and recommended LM Studio settings (temperature, top_p, max_tokens)

---

Example Quick Test Prompt (one-liner)

If result includes correctly formatted Markdown tables (HLD + Risks + Business Value) — your local setup is working.

---

Notes about images and PDF instructions

• Two sample screenshots are provided in the docs/images/ folder and referenced in INSTALL.md.
• Your initial instructions PDF (instructions.pdf) was uploaded and is included for legal/administrative reference in the repo. See the file for local-language instructions and scanned forms. :contentReference[oaicite:0]{index=0}

---

Versioning & Releases

Keep this README updated with new GGUF names when you add releases. The file list above contains all current GGUF versions and modification dates.

---

Troubleshooting

• Model download stuck in LM Studio: Check disk space and move the .gguf into a non-network path.
• Tables not rendering: Ensure you ask the model to output GitHub-flavored Markdown tables and use | Header | Header | with |---|---| separator. See PROMPTS.md for a table-enforcement snippet.
• Truncated output / token limit: Increase max_tokens in LM Studio runtime settings. Use temperature 0.0–0.3 for deterministic structured outputs.

---

Contributing

1. Fork the repo.
2. Add new GGUF under a meaningful name.
3. Update README.md table with file name and UTC mod date.
4. Submit a PR with a short note on changes & test prompts.

---

License & Credits

(Choose a license suitable for your project — e.g., Apache-2.0, MIT.)
Model training and dataset description belong to the MrMandali presales team.

---

Contact / Support

Open an issue in this repository or email the maintainer (contact in model card) for questions and commercial licensing.