Dream ★ · Two years forward
Two years of good decisions create a proposal system that prepares evidence while people keep authority. Two years of delay leave the same trusted experts carrying more comparison, compliance, and follow-up work by hand.
The Northstar, in one line
In two years, Solar365 is the commercial-and-school solar partner that an AI buying agent surfaces first and trusts most in its region — because its proposals are fast and structured, and its 3,000-install track record is machine-verifiable proof, not a homepage stat.
2 yrs
Horizon to agent-preferred commercial solar position
~8 cycles
90-day proof loops across the two-year arc
50%
Target reduction in owner time per proposal — Stage 1 proof metric
UNREAD
Timing of agent-mediated procurement — discovery watches for the first buyer signal
Good decisions compound
Each enquiry arrives with enough context to qualify. A decision packet gathers site, tariff, design, finance, delivery, and proof inputs. Mike or Austin reviews the exceptions and promise—not every transfer of data.
Weak decisions accumulate
Generic proposal software makes speed easier for every installer. If Solar365's local judgment and install record remain unstructured, faster competitors and machine-assisted buyers can compare what is legible and route around what is not.
The Northstar decomposes into five positions. Each names where Solar365 stands today and where two years of good decisions could put it. Judge whether each destination is worth choosing — the pair is the argument.
Today: Multi-segment inbound (6 ICPs); high-value commercial and school leads triaged by hand. Intake captures contact and property type only.
Two years: Commercial and school enquiries arrive into a scored queue and are routed without owner triage; the offering is machine-readable so buying agents can discover and compare it; demand is chosen, not just received.
Today: Expert judgment trapped in manual enquiry → proposal → follow-up → handoff. Binding constraint: leverage.
Two years: The enquiry-to-install loop runs as a cockpit with predictable cycle time; win/loss learning compounds into the next proposal; selling faster does not break delivery.
Today: Trust is real but human-held and unstructured — '3,000 installs,' '16% average return,' '6-year payback' live as homepage claims with no public method or machine-readable proof.
Two years: Performance is a verifiable signal — actual returns and delivery records structured as data an agent can check. The install base becomes a trust instrument, not marketing copy.
Today: No explicit capital-allocation discipline; cash reality appears late in Xero, after decisions.
Two years: Proven Stage 1/2 savings reinvested into the next bet on a named ROI threshold and kill switch; one funded growth bet (solar-farm-investor pipeline) running on real margin data.
Today: No proprietary tech; OpenSolar + Quotient + Xero + Outlook + Excel, held together manually.
Two years: One clean data spine (lead → proposal → install → performance) that every automation and every external agent reads. Not replacing the tools — ending the manual glue between them.
3,000-install track record — a trust asset waiting to compound
Solar365 has public proof of delivery at scale: 3,000 installations, 15 years of combined experience, 16% average return, and a 6-year payback claim. Most new entrants cannot match this.
The lever: Structured as machine-readable performance data — actual returns, delivery records, warranty history — the install base becomes the regional moat that an AI buying agent surfaces in every comparison.
Multi-segment demand surface with commercial priority clear
Solar365 serves six ICPs. The ICP priority order is already named: businesses first, schools second, then solar farm investors. The hierarchy exists — it just lives in memory.
The lever: A scored intake queue routes by segment value and urgency — chosen demand rather than received demand. AI surfaces the highest-value leads before expert attention is spent on lower-fit enquiries.
Expert judgment across the full proposal loop
Mike and Austin carry site assessment heuristics, pricing logic, delivery knowledge, and supplier relationships built over years. That judgment is the differentiator.
The lever: Codify the inputs behind the judgment — segment type, site facts, ROI confidence, delivery risk — into a decision brief the system prepares before expert review. Judgment stays human; brief-building becomes a system.
Existing tool stack that can support Stage 1
OpenSolar, Quotient, Xero, Outlook, and Excel are already in daily use. No greenfield build is required for cycle 1.
The lever: Stage 1 does not replace the tools — it ends the manual glue between them. The data spine routes from OpenSolar assumptions into Quotient output without rekeying.
Decision speed — small-team structure
No board approval cycles, no shareholder optics. Mike and Austin can commit to a bounded discovery sprint within days once GO conditions are met.
The lever: Speed-to-pilot is a structural advantage over larger installers with procurement committees. The 14-day discovery window can confirm GO/NO-GO before competitors have finished a steering meeting.
Agentic procurement routes past slow respondersUNREAD — date from observed buyer behaviour
If commercial or school buyers begin using agents to compare quotes, returns, and payback, slow or unverifiable suppliers become harder to shortlist. The direction is plausible; the closing horizon is not yet evidenced.
Leading indicator: The first major commercial buyer who asks for machine-readable performance proof before shortlisting Solar365 is the leading indicator. By the time that request is routine, the window to adapt has closed.
Generic AI proposal tools commoditise response speedNow–18 months
Generic AI tools let any competitor produce a fast commercial solar quote. Speed alone is no longer a differentiator if every entrant can match it.
Leading indicator: When competitors produce AI-assisted proposals and Solar365's process remains manual, any response-speed advantage disappears — and the install-record moat has not yet been structurally built.
Delivery capacity cannot absorb faster salesConcurrent with Stage 1
Faster proposals create more accepted work. If subcontractor readiness, installer capacity, and equipment lead times are not visible before commercial promises are made, a faster proposal loop damages the 3,000-install trust record faster than it was built.
Leading indicator: Kill signal in transformation-plan.md: 'Delivery stress — faster sales creates unowned subcontractor or install handoff risk.' Stage 1 must not increase sales speed beyond visible delivery capacity.
Vendor-led transformation trapNow
The most likely failure mode is engaging AI tooling on a vendor's terms — scope expansion, seat licensing, and promise inflation that does not map to Solar365's actual constraint (proposal leverage).
Leading indicator: The bounded-bet model — GO conditions confirmed before build, kill signal at Week 6, 14-day discovery before commitment — is the structural protection. Proof before platform.
Resilience read
Hardest force: Buyer power ↑↑. If commercial and school buyers begin using agents to compare quotes, returns, and payback, slow or unverifiable suppliers become harder to shortlist. The direction is plausible; the arrival date is unknown.
How the position absorbs it: The two-year position combines machine-readable performance proof with scored, structured proposals whose speed is measured against Solar365's baseline. The first buyer request for machine-readable proof is the adaptation trigger.
Second force: Substitutes / new entrants ↑. Generic AI proposal tools let anyone produce a fast quote. The defence is not speed alone — it is speed plus the verifiable 3,000-install track record and local delivery capability a generic entrant cannot fake. The data moat is the answer to commoditised speed.
The Dream is not a prediction; it is a position with instruments. These four signals say whether the future is arriving on schedule — and each names why it matters.
Proposal response speed vs competitors
What to read: Time from commercial or school enquiry to decision-ready proposal sent, compared to nearest competitors.
Why it matters: Agent-mediated procurement routes to whoever responds fastest with verifiable numbers. A 2-day lag becomes a structural disadvantage when agents compare in seconds.
Agentic procurement arrival rate
What to read: How many major commercial or school buyers are demanding machine-readable quotes or performance proof before shortlisting.
Why it matters: Buyer power is the hardest force. The first buyer who requires structured data is the leading indicator. Waiting until the second buyer is late.
Owner time per proposal (leverage ratio)
What to read: Expert hours from qualified enquiry to proposal sent, by ICP segment and proposal type.
Why it matters: The binding constraint is leverage, not market fit. This is the primary Stage 1 proof signal — if it does not drop 50% in 90 days, Stage 2 is not funded.
Generic AI proposal tool adoption by competitors
What to read: How many NZ commercial solar installers are producing fast, AI-assisted proposals and whether output quality is comparable.
Why it matters: Speed is necessary but not sufficient. The defence is verifiable track record plus local delivery capability — assets a generic entrant cannot copy.
From a leverage-trapped installer doing high-value work by hand, to a learning commercial system: expert judgment leveraged across the whole pipeline, commercial/school/solar-farm-investor segments compounding, and a 3,000-install performance dataset that becomes a regional moat — the agent-preferred commercial solar supplier in its market. Reputation stops being a claim and becomes machine-checkable proof that compounds with every install.