Umbrella for higher-dollar, sporadic engagements beyond the agency. First vertical: on-prem AI for law firms (~$100K engagements); the fix-it entry point ("your AI is broken, we know why") may close faster than greenfield. Methodology is engagement-ready: six sellable modules, Foundation/Premium tiers, three named differentiators (from the RAG training sessions). No legal entity yet — formalize when the first deal is imminent. Related-but-separate: the Eric mentorship (drawer 13) provides accountability/timing counsel on ventures like the boat-financing SaaS.
Non-Negotiables
Client-facing rules — no engagement ships without theseDocument taxonomy + metadata extraction are enforced at ingestion. No exceptions.
WhyEmbedding models encode semantic similarity, not authority or recency. Only metadata distinguishes a 2015 draft memo from a current policy.
Metadata is not optional — it is the foundation of retrieval quality.
WhyWithout it, the retriever can't filter by status, version, access level, or authority rank.
Systems deployed without taxonomy and metadata are production liabilities.
WhyThey produce inconsistent answers that erode partner trust and fail audit.
Ongoing curation and self-healing maintenance are the product, not add-ons.
WhyDay-one effectiveness only stays day-one-thousand effectiveness if the knowledge base is treated as a living system.
Sellable Methodology — the modules a client engagement follows
Document Taxonomy
Four-folder pattern (Policies / Templates / Case Files / Correspondence) with enforced metadata schema. For 50K-doc corpora, discover categories by sampling ~5K with an LLM and getting owner approval.
Precedence Rules
Newer > older, policy > opinion, verified > draft. Implemented in Qdrant via priority-then-date ranking. Makes answers consistent, auditable, defensible.
Structured vs. Unstructured Ingestion
Manual markdown curation for the critical 100–500 documents. Automated ingestion for the 49,500+ rest. 80% retrieval quality at 20% of the curation cost.
Self-Healing Maintenance
Three-layer monitoring: explicit feedback + implicit behavioral-signal detection + autonomous audit cycle on a fine-tuned domain LLM. This is the retainer engine.
System Prompt Engineering
Role + authority, citation + traceability, confidence gating. Plus Intelligent Query Gating: recognizes vague prompts, gates with 2–3 clickable clarifications, then teaches the user the prompt template at end of session.
Testing and Validation
Golden question sets (50+ min) curated by firm SMEs. Precision@K, Recall, MRR, hallucination rate. Monthly rerun tracks drift; quarterly audit of underperforming sections.
Differentiators
What we do that ship-and-leave competitors don'tSelf-Healing RAG
Three layers — explicit feedback + implicit signal detection + autonomous audit. Most competitors ship-and-leave; we read the room and fix the knowledge base before the user realizes there was a problem.
Intelligent Query Gating
Users get real-time guided clarification + end-of-session prompt templates. Within weeks, they ask better questions, the tool 'works better' — your AI reputation is protected.
Retainer-First Architecture
Build in month one. Operate and improve every month after. What you knew about your system on day one still holds on day one thousand — because we treat it as a living system, not a one-time deployment.
Service Tiers
Foundation
RAG deployment, complete
~$100K initial + monthly retainer
- • Full six-module RAG deployment playbook
- • On-prem Qdrant + embedding model + local LLM
- • Taxonomy + metadata + precedence rules enforced at ingestion
- • Structured/unstructured hybrid curation
- • Self-healing three-layer monitoring
- • Monthly golden-question validation + retainer
Best for Most clients. Fast to deploy (weeks, not months). Easy to maintain (re-embed when docs change). Full source attribution on every answer. Scales without retraining.
Premium
PremiumFoundation + quarterly fine-tuning
~$300K enterprise tier from the 5-tier pricing
- • Everything in Foundation
- • Quarterly fine-tuning of the underlying LLM on client documents
- • Model bakes in domain language, terminology, tone
- • RAG still handles retrieval + source attribution
- • Fine-tuned model 'thinks' like the firm before retrieval runs
- • On-prem fine-tune jobs run over a weekend on M3 Ultra with 10K-doc examples
Best for Clients who say 'our attorneys have a house style we always rewrite drafts to' or 'partners want the AI to sound like us.' Foundation handles lookup; Premium handles on-brand generation.
Pitch-Ready Language
Paste-ready executive linesThe Nonverbal Cues Analogy
Executive pitch opener, website hero, keynoteEvery executive has been through sales training. A prospect says 'I'm interested,' but their arms are crossed, they're leaning back, they're not asking questions — the nonverbal cues tell you the real story. Your knowledge base has the same dynamic. Users won't always say 'that answer was wrong,' but their behavior leaks the truth. They rephrase the question. They abandon the conversation. Most RAG systems ignore these digital nonverbal cues. We don't. We read the room and fix the knowledge base before the user realizes there was a problem.
The Retainer ROI Line
Pricing conversation, sustainability argumentThe cost of maintaining and fine-tuning this platform will pay for itself one hundred times over. What you knew about your system on day one can still be true on day one thousand — but only if you treat it as a living system, not a one-time deployment. We make that easy.
The CTO Value Prop
CTO / CIO pitches where adoption has already burned themCTOs have invested millions in LLM infrastructure. Users blame the tools when they don't know how to use them. Bad prompts → bad answers → users think the tool is broken → they badmouth it internally → your AI reputation tanks, even though the tool is fine. Our system solves this silently. Guided clarification in real time. Users learn prompt patterns. Within weeks they're asking better questions. The tool 'works better' — but really your users got smarter. You protect the investment. You protect your reputation.
The Institutional-Knowledge Line
When the client thinks they just need searchWe're not just indexing your documents. We're encoding your firm's institutional knowledge hierarchy into the retrieval system.
The Defensibility Line
Legal / compliance / regulated-industry pitchesWithout precedence rules, your system is unpredictable. Same query, different day, might pull different sources. With precedence rules, it's consistent, auditable, defensible. Partners trust it. They cite it in client work. It becomes an asset.
The 80/20 Line
When cost-of-curation is the objectionWe invest curation effort where it matters most. Your active policies and templates get structured, human-reviewed markdown. We automate the rest. Eighty percent retrieval quality at twenty percent of the curation cost.
The Retainer Close
Closing the monthly-retainer commitmentEvery month we run your test suite. We report metrics. We flag degradation. We recommend audits. That's continuous assurance that your knowledge base stays production-ready.
Training Curriculum
Voice-chat tutoring with Claude Chat, one topic per ~25-30 min session. Currency enforced (today's date + web search from turn 1). Each session produces a journal transcript; sellable outputs distilled into methodology/.
✓ Completed
RAG Fundamentals
2026-04-21 (captured)Vocabulary (embeddings, vectors, chunking, vector DBs), full pipeline walkthrough, five failure modes, RAG vs. fine-tuning. Origin of the Premium tier idea — Topher's insight: 'Why wouldn't you set up RAG AND fine-tune so the model leverages the RAG platform?'
RAG Context Architecture
2026-04-20Full six-module sellable playbook, four non-negotiables, three differentiators, 20-attorney / 50K-document capstone. Nonverbal cues analogy, CTO value prop, ROI line. Session hit token limit mid-capstone critique.
⏳ Backlog
Active Items
Synced from the Kanban boardWaiting on You
(2)Demo buildable on M4
Weekend build per the 16-section brief, when prioritized.
Fold Foundation/Premium tiers into the on-prem brief
/info/aspire-ventures-onprem-ai predates the premium-tier concept.
In Progress
(1)AI Readiness Assessment entry product
$15-25K, 2-3 weeks; low risk, upsells naturally into Foundation.
Blocked
(3)Fix-it positioning + first engagement
⚡ HIGHPipeline empty — no prospects engaged. Gate per doctrine: Aspire Digital internal setup FIRST; ventures stay behind the agency gate.
Law firm target identification
Abstract — no specific firm targeted yet.
Legal entity formalization
When the first engagement is live, not before.
Key People
Eric Fatkin
Mentor — accountability + transition timing (own drawer, #13); NOT a finance seat
Recent Events & Decisions
TCYM Grace v3 assets landed in this drawer
Single-agent demo redesign ratified — Grace answers as fictitious Harmony Funeral Home. KB + prompt live in tcym/ (harmony-funeral-home-kb.md, grace-demo-prompt.md); paste kit at /info/tcym-grace-demo. TCYM program itself is tracked in the Aspire Digital drawer; these are the venture-side source files.
Boat-financing SaaS — meeting #3 update captured
Working through the Eric mentorship channel; update in boat-financing-saas-meeting-3-update.md.
Funeral front-desk play evaluated
plays_comparison + play_funeral_front_desk — later retired into Aspire Concierge (Aspire Digital drawer owns the funeral vertical now).
Methodology + pitch copy shipped
Training 2.1 (RAG Fundamentals) folded in; 7 paste-ready pitch lines with placement guide.
RAG Context Architecture (2.3) — six-module playbook
Full sellable playbook, four non-negotiables, three differentiators, 20-attorney/50K-doc capstone.
Memory Files (source of truth)
The markdown files Aria reads for deeper contextmemory/drawers/ventures/boat-financing-saas-meeting-3-update.md 2026-05-08 update to the marine dealer financing SaaS thesis. Eric read the brief, validated the opportunity, but research revealed several factual errors and structural changes (Bank of the West dead, Trident ownership, Synchrony cap) that need to land in v2 of the brief.
memory/drawers/ventures/clients.md memory/drawers/ventures/overview.md Moonshot consulting — on-prem AI for law firms, payments+AI enterprise consulting, AI implementation fixing
memory/drawers/ventures/play_funeral_front_desk.md Productized HighLevel-leveraged service for independent funeral homes. Eternal Front Desk (or final brand TBD by Jaime) — Voice AI receptionist + preneed nurture + reputation. $2,500 setup + $797/mo. 90-day target $55-75K, 120-150 day target $100K. Separate from RAG/law-firm Ventures track.
memory/drawers/ventures/plays_comparison_2026-04-27.md Side-by-side ranking of 10 productized HighLevel-leveraged plays scoped 2026-04-27. Funeral homes locked at
memory/drawers/ventures/methodology/ Sellable methodology distillations (rag-deployment-playbook, pitch-copy, service-tiers)
memory/drawers/ventures/journal/ Training session transcripts + INDEX
memory/drawers/ventures/efd/ EFD archive (retired into Aspire Concierge)