Instrument 03 · The operating spine
The best stack is not the one with the most software. It is the one where every lead, design, promise, invoice, and result has one owner—and every handoff can fail visibly.
Best-fit verdict
Do not replace the current stack. First prove the qualified-enquiry → OpenSolar-project handoff. Keep a manual review gate until its field coverage, failure rate, and per-project cost are known.
Current reading
UNREAD — the tools are named, but field ownership, enabled integrations, duplicate entry, failure alerts, and New Zealand API costs are not yet mapped.
OpenSolar designs, Quotient presents, Xero accounts, and Stackmates qualifies and learns. Integration earns its place one measured handoff at a time.Each layer gets one job. A product may help elsewhere, but it does not become the source of truth there.
Stackmates CRM-004 + current website intake
Who the buyer is, which segment fits, what is missing, and whether expert attention is warranted.
A qualified lead may create a reviewed OpenSolar project. An unqualified lead may not.
OpenSolar
Site, system, design, production, hardware, and solar pricing assumptions.
Use its API or supported connector only after the required fields and per-project charge pass the trial.
Quotient now · Stackmates quote engine behind a gate
The versioned commercial offer the buyer sees and accepts.
Quotient stays the live surface. Stackmates QUOT-001 is Built but has a route gap; compose it only after a proposal-path proof.
Xero
Customer, invoice, payment, tax, and accountant-reviewed financial records.
Reuse Quotient's accepted-quote → Xero invoice integration before writing a custom accounting path.
Stackmates evidence register + governed workflow
Why a lead qualified, what was promised, what changed, and what the delivered system actually achieved.
Start as a reviewed register. Automate only fields that have an owner, an alert, and a replay path.
Capability state matters. Built means the engineering exists—not that it is already configured or live inside Solar365.
Score and route enquiries using Solar365's own segment, fit, missing-data, and attention rules.
Built use case; Solar365 rules and workflow still require discovery.
Compose a structured quote from proposal evidence while keeping the commercial promise reviewable.
Backend proof passes 6/6; the route and sellable UI remain a gate.
Parse, extract, index, and retrieve from proposal or procurement documents when the first workflow actually needs it.
Proven on a real PDF; currently RFP-focused, without general OCR or full-text coverage.
OpenSolar field mapping, failure recovery, delivery-capacity checks, and post-install outcome evidence are Solar365-specific work.
Not sold as existing capability. Discovery decides compose, demand, or no-build.
Choose tools by the failure they remove, not by feature count.
Solar design and proposal inputs need one specialist home.
External API access is per-project; connectors are separately priced by region.
Accepted quotes already need a reliable accounting handoff.
Confirm the live integration and field ownership before adding another quote or invoice system.
One measured handoff removes enough re-entry or delay to pay for access.
Trial exact field coverage, wallet behaviour, failure alerts, replay, and NZ pricing.
A small event handoff is stable and does not own business truth.
Automation without an owner and recovery path creates invisible operational debt.
OpenSolar fails a named design, proposal, API, support, or cost requirement in the trial.
Vendor comparisons are claims, not migration evidence. Price data migration and retraining before replacing anything.
Prove one transfer. Do not automate the whole lead-to-cash path.
Approve the 14-day data-flow test—not a platform replacement. Its output is a field map, capability verdict, costed integration choice, and GO / CHANGE / STOP decision.