Every move, in one place.

Grounding, client confidentiality, and what's never safe to paste. The skeleton, practiced to reflex.

Retail and office leases plus investment sales: abstract a lease, value on the income approach, and run the investor's numbers.

Design a repeatable task into an AI workflow. With a human approval gate exactly where it belongs.

One platform, mapped to the job: a Project for your standing context, a Cowork skill for a repeatable prompt, MCP connectors for your live inbox and calendar.

Compare spaces, deliver the BOV, audit the CAM, catch the estoppel mismatch, and run the PSA and 1031 clocks, the diligence half of commercial.

Commercial deals III, the class-specific reads: multifamily loss-to-lease, industrial spec fit, sublease deal-killers, retail breakpoints and co-tenancy, and the three-seat risk brief. Grounded, cited, nothing invented.

Owner canvassing and outreach for ten-year hold periods: trace the LLC through the public record, write the letter about THEIR building, canvass so it compounds, hold a two-year cadence without inventing a touch, prep the call, debrief it clean.

The commercial operator's desk: the lease's critical-date register, the month's variance narrative without an invented cause, the renewal that lands, vendor bids leveled on real scope, the asset's annual business plan, and tenant notices that commit to nothing you can't back.

The lender's side: size the loan to its binding constraint with the mortgage math in a real calculator, draft the credit memo without recomputing a ratio, foot the sources-and-uses, read covenants into a compliance checklist, QC the package before credit, and write the watchlist flag a servicer can defend.

Raising and reporting to investors under securities discipline: the distribution notice that states never promises, the LP update that survives scrutiny, the waterfall read from the agreement, the DDQ answered from real facts, the raise one-pager that labels every projection a target, and the deck red-teamed before a regulator does it.

Two report-heavy desks under strict grounding: the USPAP scope structured without stating a value, the sales-comparison grid built from comps you can defend, the zoning code read without hallucinating it, the appraisal red-teamed for credibility gaps, the entitlement path mapped from the pasted process, and the development pro-forma skeleton built without inventing a number.

The land developer's and land flipper's desk under strict grounding: raw land screened against your program, valued from your comps as a range, run through the deal-killer scorecard, its chain of title traced, its split screened, the max price and seller-carry terms set up without ever inventing a number, an approval, or an appraisal.

CRE-institutional market work under strict grounding: net absorption and vacancy quoted to the report, an asking-vs-effective rent survey, a cap-rate trend that shows its comps, the deliveries pipeline in a matrix, a competitive set built only from the comps you pasted, and a submarket investment thesis where every figure is sourced or marked UNKNOWN.

The borrower's side of the loan under strict grounding: the financing package built from pasted facts, the deal screened against each lender's stated box, a term sheet marked up without inventing market leverage, live quotes turned into a client debt-options memo, and the rate-lock decision briefed. Every number quoted, the compound math handed off, never a rate forecast.

The landlord's side of leasing under strict grounding: a value-framed leasing proposal with asking rent left as your number, the stacking plan and rollover schedule read from the rent roll, a lease-comp survey positioning your space on effective rent, the owner's renew-vs-release math set up not guessed, a marketing package true to the pasted specs, and a win-the-listing pitch red-teamed of every claim it can't back.

The close under funds-and-legal grounding: build the deliverables matrix from the contract, read the escrow instructions to their conditions, tie every settlement-statement line to its source without recomputing a proration, and track the lender's conditions to fund, all quoted, nothing authorized, and never a 'clear to close.'

The land flipper's front-end deal engine under strict grounding and compliance: a deal thesis turned into filter criteria (never owner facts), a list scrubbed and clustered with every flag a hypothesis, a transparent motivation score that never touches a protected class, outreach screened for TCPA and consent (never cleared to send), a marketplace mispricing scan, and listing copy true to the parcel. Targeting by behavior, never a person; legal calls routed to counsel.

Audit the deal week before you delegate a line, give every figure a source cell before it reaches the memo, cost the lease-abstract run, wrap the day into the deal log, and know what binds an AI-drafted BOV.