How an engagement actually runs.
Here is what the work looks like from the inside. Six examples, each drawn end to end: where an expert decides,
what runs automatically, and which specialist does which piece. Each one adds to a shared knowledge base,
so a rule or a cost researched for one building is already on file for the next.
Who is involved
Expert
Your Soapbox expert, in the thread with you. The expert sets the assumptions, confirms what the work finds, settles any conflict, and hands you the report.
Lead agent
One per engagement. It runs the conversation, decides the order of work, and assembles the report. It never produces a number itself.
Specialists
One job each. They work from shared files, hand back what they found, and stop when something is missing rather than guess.
- Expert decision
- Automatic check
- Lead agent
- Specialist
- Repeats per building
- On site
The six examples
One building 7 specialists From a measured baseline to a phased capital plan with the returns attached.
The expert decides twice, because this plan commits capital. Once when the baseline and carbon target are set, and again before the measure list becomes a roadmap.
One building 8 specialists We run the audit ourselves, including the site visit.
The expert decides twice: once to confirm the baseline and the survey before measures are proposed, and once to approve the combined energy and water roadmap.
A pre-committee screen from an offering memorandum, or from just an address.
Nobody decides mid-run. This is a screen, and speed matters more than precision, so it runs start to finish and comes back to the expert once.
A portfolio with almost no data 7 specialists Start from addresses and who pays the utilities. End with a ranked list of the buildings worth asking for documents on.
The expert decides once, at the start: which jurisdictions matter, which measures are off the table, and how many buildings you can request documents on. The rest is ranking.
Tens to hundreds of buildings 6 specialists Screen a whole portfolio with real data on every building, in one pass.
The expert decides once, at the start: the building list, the rates, and how carbon links to value. After that the run is automatic, because the value of a portfolio pass is that it finishes.
A portfolio under offer 3 specialists A full plan for every building, with the expert's decisions made once per building type and applied across the rest.
The expert decides twice, once per building type rather than once per building: first on the representative buildings, then on any building that turns out not to fit its group.
What one engagement learns, every engagement keeps
Every engagement raises questions nobody has answered yet. Rather than guess, or make you wait, the work spins
them out into a separate research track. Whatever comes back is checked, made general, and saved for every
future engagement, not just the one that asked.
A gap surfaces
Mid-engagement a specialist hits something we do not hold: an incentive program we have not seen, a local rule, a cost with no source. It does not guess. It files the question and carries on.
Research runs separately
The question leaves the engagement and runs on its own clock. A research agent works it from primary sources and cites each one.
Checked before it is trusted
Each finding is checked back against the document it claims to come from. Anything uncertain is rejected rather than softened.
Made general, then saved
Anything specific to your portfolio is stripped out. What remains is a general fact, a rate, a rule or a cost band, saved to the shared knowledge base.
The next engagement starts knowing it
A building performance rule researched for one asset is already on file for the next asset in that city. Nobody researches it twice.
Nothing crosses between organizations. Your figures, documents and portfolios stay inside your workspace. Only the
general finding is shared forward.
Run one on your portfolio.
Tell us what you need answered and we will pick the right approach.