Trinity Graph, Annotated Business Model Canvas, CFO Persona Architecture & System Architecture — Session 7 Midterm
Founded St. Patrick's Day 1970 · Rockland/Hanover, MA · $100–150M Est. Revenue
Buckley Associates is one of the Northeast's leading suppliers of commercial HVAC equipment and solutions, operating a hybrid business model across three distinct revenue pillars — each with its own cost structure, margin profile, and data complexity.
| Pillar | Description | Margin Complexity |
|---|---|---|
| Distribution / Rep Sales | 3,800+ stocked SKUs across 5 warehouses, 30+ manufacturer lines (Greenheck, Fujitsu, Price Industries, Big Ass Fans, etc.), free next-day delivery | Thin margins, high volume. Hidden logistics cost in "free delivery." Margin leaks in freight, returns, inventory carrying cost across 5 warehouses. |
| In-House Manufacturing | 50,000 sq ft union sheet metal shop. Roof curbs, equipment supports, flexible duct, fire dampers, louvers, fittings. | Labor-intensive, union economics. Margin depends on estimating accuracy vs. actual shop floor execution. Classic E2/JobBOSS territory. |
| Services | Equipment startup, commissioning, troubleshooting, project estimation, design assistance, education & training | High-margin but hard to track. Often bundled or given away to win equipment deals. True profitability invisible. |
| Location | Function |
|---|---|
| Hanover, MA (385 King St) | Warehouse / Manufacturing HQ (50K sq ft shop) |
| Rockland, MA (1099 Hingham St) | Corporate Office |
| Newington, CT (15 Progress Circle) | Regional Office / Warehouse |
| Milford, CT (294 Quarry Rd) | Regional Office / Warehouse |
| Albany, NY (120 Railroad Ave) | Regional Office / Warehouse |
| Manchester, NH (55 Buckley Circle) | Regional Office / Warehouse |
| Gorham, ME (510 Main St) | Regional Office / Warehouse |
Greenheck (fans, ventilators, dampers, louvers, kitchen ventilation, lab exhaust) · Price Industries (air distribution, terminal units, sound control, critical environments, sustainable design) · Fujitsu (VRF, mini-splits) · Big Ass Fans (HVLS) · Air Concepts · AQC Industries · DuctSox · Eastern Sheet Metal · Aldes (energy recovery, airflow controls) · SolutionAir · Systemair · Filtration Group · GPS Air (NPBI) · Indeeco · NovelAire · Car-Mon · Continental Fan · Powered Aire · Reversomatic · Thermolec · Lumalier · Delta · fanAm · Climetec · Trion · Sternvent · Van Packer · Young Regulator · Jeremias · Pro-R Duct · Airius · Interzon
Social (WHO) · Knowledge (WHAT) · Generative (WHAT IF)
The structured intelligence Enginuity builds from Buckley's raw data — the moat that Google, Tableau, or any competitor cannot replicate from outside the firewall.
Engineering builds a quote with estimated labor hours, material costs, and margin. The shop floor executes the job with actual labor, actual material waste, and actual timeline. The gap between those two numbers is where margin dies — and Buckley's current systems don't surface it until month-end close when it's too late. Enginuity connects quoted estimates to actual execution in real-time, job by job.
Buckley generates mountains of data every day across 7 locations, 3,800+ SKUs, a union shop, and a service operation. But the data is dirty — rounded timecards, inconsistent SKU naming across locations, manual spreadsheet entries, duplicated vendor codes. Tableau visualizes whatever it's fed. Enginuity quarantines the noise, identifies systematic data quality issues, and only promotes clean, trustworthy data to the insight layer.
Sales, Engineering, Manufacturing, Distribution, and Finance each use different identifiers for the same entities. A roof curb might be "Custom RC-12" in Engineering, "Job #4822" in Manufacturing, "PO-7741" in Purchasing, and "Revenue Line 3.2.1" in Finance. Enginuity maps these siloed identifiers into a single product/project lifecycle — creating cross-functional visibility that doesn't exist today.
Over time, Enginuity builds customer-level and project-type-level margin risk profiles. It learns that certain contractors always request mid-project changes, certain product configurations always exceed estimated labor, and certain locations consistently over/under-perform. This predictive layer informs future quoting and flags at-risk jobs before they ship.
The CFO's 5-year plan requires forward-looking data: which revenue pillar is growing, which is margin-dilutive, where to invest in capacity, when to hire, whether to open an 8th location. Today these decisions are informed by backward-looking Tableau dashboards. Enginuity transforms historical pattern recognition into predictive strategic modeling.
Whatever source systems feed Tableau today (QuickBooks, Sage, spreadsheets, or a legacy ERP), Enginuity doesn't replace them. It sits on top, ingesting their outputs via CSV/API, cleansing and structuring the data, and creating the intelligence layer that these systems were never designed to provide.
| CSV Stream | Source | What It Contains | Insight Unlocked |
|---|---|---|---|
| Quotes / Estimates | Engineering / Sales | Estimated labor, materials, markup, delivery timeline per project/job | Theoretical margin — what the company expects to earn |
| Execution / Production | Shop Floor / Ops | Actual labor hours, material consumption, scrap, rework, production timeline | Actual margin — what the company really earned. Variance = the reality gap |
| Purchasing / Inventory | Procurement / Distribution | PO data, vendor pricing, inventory levels across 5 warehouses, freight costs | True COGS, carrying costs, logistics burden of "free delivery" |
| Finance / Revenue | Accounting / CFO | Invoiced revenue, AR aging, payment terms, revenue by pillar/location/customer | Cash flow reality, customer profitability, pillar-level P&L |
The intelligence agents that transform raw knowledge into actionable business recommendations — starting at the CFO level and trickling down to every functional team.
Aggregates insights from all downstream agents. Generates: consolidated P&L visibility across all 3 pillars, 5-year plan progress tracking, capital allocation recommendations, M&A readiness scoring, and strategic business recommendations. Outputs: weekly executive brief, monthly strategic review deck, quarterly board-ready financial narrative.
Feeds CFO Agent with: Quote accuracy trending, systematic estimation errors by product type, recommended labor rate adjustments based on actual shop floor data. Generates: Estimation confidence scores per job, automated quote audits, "this job type always runs 22% over estimated labor" alerts. Reduces the reality gap at the source.
Feeds CFO Agent with: Revenue attribution by marketing channel, manufacturer co-op fund utilization, training/education event ROI, customer acquisition cost by segment. Generates: Budget reallocation recommendations, manufacturer line performance rankings (which of the 30+ lines actually drive profit vs. just revenue), campaign effectiveness scoring.
Feeds CFO Agent with: Win/loss rates by customer and product type, pricing elasticity analysis, volume vs. margin tradeoffs, customer lifetime value. Generates: Dynamic pricing recommendations grounded in actual execution costs, customer profitability rankings, "which deals should we walk away from" analysis, volume discount optimization across manufacturer rebate tiers.
Feeds CFO Agent with: True delivery cost per order by location, inventory carrying cost across 5 warehouses, SKU velocity analysis (which of 3,800+ items actually move). Generates: Warehouse consolidation recommendations, "free delivery" profitability threshold analysis, dead stock identification, optimal stocking level recommendations per location.
Feeds CFO Agent with: Labor variance by job type, union labor utilization rates, material waste trending, production bottleneck identification, capacity utilization. Generates: Shift optimization recommendations, "make vs. buy" analysis for specific product lines, labor cost forecasting, preventive quality alerts when a job is trending toward margin erosion mid-production.
| Category | What It Generates | Who Benefits |
|---|---|---|
| Pricing Intelligence | Margin-optimized pricing based on actual cost data, not theoretical estimates | Sales, CFO |
| Marketing ROI | Which channels, events, and manufacturer relationships actually drive profitable revenue | Marketing, CFO |
| Sales Distribution | Geographic and customer-segment analysis — where to invest sales resources | Sales, CFO |
| Volume vs. Margin | Which customers and product lines deliver volume without margin — and vice versa | Sales, CFO, Ops |
| M&A Due Diligence | Buckley's own financial data structured for acquirer-grade transparency — or applied to evaluate acquisition targets | CFO, Board |
| Strategic Recommendations | Data-driven input to 5-year plan: expand which pillar, add which location, invest in which product line | CFO, Leadership |
Every CSV processed, every job completed, every quote vs. actual comparison adds another data point to the Knowledge Graph. The Generative agents get smarter with each cycle. After 6 months of deployment at Buckley, the system knows things about this business that no consultant, no ERP vendor, and no outside AI could ever replicate. That compounding intelligence IS the moat.
9 BMC boxes mapped to Trinity Graph layers — grounded in Buckley's real business
Role-based (not person-specific) — universally applicable to any hybrid manufacturer CFO
$100–150M revenue · 3 revenue pillars · 7 locations · 50–200 employees · Legacy IT stack
Owns the P&L across all revenue streams. Owns the 5-year strategic plan. Reports to ownership/board. Makes capital allocation decisions (new locations, equipment, headcount, M&A). Responsible for financial reporting, cash flow management, and operational efficiency metrics. The person who gets asked "why did margins drop?" and currently doesn't have a reliable answer.
| Layer | Current Tool | The Problem |
|---|---|---|
| Visualization | Tableau | Shows what you feed it — doesn't challenge whether the inputs are true |
| Source Systems | Unknown (likely QuickBooks, Sage, spreadsheets, or legacy ERP) | Dirty data, inconsistent entry, no cross-system reconciliation |
| Intelligence Layer | DOES NOT EXIST | This is the gap. This is where Enginuity lives. |
Drowning in data from 3 revenue streams, 7 locations, 30+ vendor lines, and 3,800+ SKUs. No single source of truth. Every report requires manual reconciliation.
Cannot see real-time margin by job, customer, product line, or location. Discovers margin erosion 30+ days after the fact during month-end close.
Tableau dashboards look professional but nobody audits what feeds them. Decisions are being made on data that may include rounded timecards, inconsistent SKU codes, and duplicated entries.
Strategic decisions (open new location? invest in manufacturing capacity? pursue M&A?) require forward-looking data. All current tools are backward-looking. The 5-year plan is built on gut feel and outdated spreadsheets.
Sales, Engineering, Ops, and Finance each have their own data, their own definitions, and their own version of reality. The CFO is the only person who needs to reconcile all four — and has no tool to do it.
"Free next-day delivery" across 5 warehouses. Service work bundled into equipment deals. Union labor allocated to jobs that weren't properly estimated. The true cost of doing business is buried.
One unified view across all 3 revenue pillars, all 7 locations, all 30+ vendor lines. Data cleansed, reconciled, and trustworthy before it reaches any dashboard.
Margin tracked at the job, customer, product line, and location level — in real time, not 30 days later. Alerts when a job is trending toward margin erosion mid-execution.
Forward-looking models that inform the 5-year plan: which pillar to grow, where to invest, which customers to prioritize, when to hire, whether to pursue M&A.
Every number the CFO sees has been audited by the data health layer. No more gut feel. No more "I think margins are fine." Data-driven certainty that survives board-level scrutiny.
| Objection | Response |
|---|---|
| "We already have Tableau" | Tableau is your mirror. Enginuity audits what's feeding the mirror. We make Tableau trustworthy — we don't replace it. |
| "Our data is too messy for AI" | That's exactly why you need us. Messy data is our starting point, not our blocker. The data health layer was designed for this. |
| "We can't afford another system" | You can't afford not to know where your margins are leaking. One bad job you catch early pays for a year of the platform. |
| "IT will never approve this" | We work with CSV exports from your existing systems. No API integrations required. No ERP modifications. IT doesn't need to change anything. |
| "How do I know the AI's insights are right?" | Every insight traces back to your source data with full audit trail. We quarantine suspicious data points and flag them — you decide what's true. |
Where Enginuity sits relative to Buckley's existing stack
Strategic intelligence, 5-year plan modeling, M&A readiness, consolidated P&L across all 3 pillars
Powered by CFO Strategic Agent — aggregates all downstream agent insights
6 functional agents: Engineering/Estimation · Marketing · Sales/Pricing · Distribution/Logistics · Manufacturing · CFO Strategic
Each agent generates role-specific insights → all feed up to CFO Master Agent
Entity resolution · Silo translation · Margin variance modeling · Customer risk profiles · Predictive intelligence
The moat — behind-the-firewall ground truth that compounds with every data cycle
Data cleansing · Quarantine engine · Anomaly detection · Source validation · Audit trail
Catches dirty data BEFORE it reaches insights — rounded timecards, duplicate entries, inconsistent codes
4 data streams: Quotes/Estimates · Execution/Production · Purchasing/Inventory · Finance/Revenue
Simple CSV exports from existing systems — no API integration required
Legacy ERP / QuickBooks / Sage / Spreadsheets → Tableau (visualization only)
Enginuity does NOT replace these systems. It sits on top. IT changes nothing.
Buckley's current architecture has two layers: source systems (Layer 0) and Tableau (which currently sits where Layer 5 is). Layers 1 through 4 do not exist today. That's four layers of intelligence — data ingestion, data health, knowledge graph, and generative agents — that Enginuity provides. Tableau can stay exactly where it is. It just gets fed clean, trustworthy, insight-rich data instead of raw, unaudited noise.
| Capability | Tableau (Today) | Enginuity (New) |
|---|---|---|
| Data Visualization | ✅ Excellent | Not the focus — let Tableau do this |
| Data Cleansing | ❌ None | ✅ Automated quarantine & validation |
| Cross-Silo Entity Resolution | ❌ None | ✅ Maps identifiers across departments |
| Margin Variance Analysis | ❌ Can display if manually built | ✅ Automated quote vs. actual comparison |
| Predictive Intelligence | ❌ Backward-looking only | ✅ Forward-looking margin & risk models |
| Strategic Recommendations | ❌ None — it's a mirror | ✅ Agent-generated business insights |
| Data Integrity Audit | ❌ Trusts whatever it's fed | ✅ Questions everything before promoting it |