Intake, decision, blueprint, execution, and end-state design.
Nothing below is running in production. These five layers are the architecture PolarAIs is designed to deliver, described as designed rather than as shipped. The prototypes linked from each layer are clickable and live.
Every entity enters as a structured fingerprint, not a folder of PDFs: deal, footprint, model, people, tech, and compliance captured as comparable fields the moment diligence closes. On Day 1 you hold one scoring-ready profile of every entity instead of chasing status across legal, tax, and IT for six weeks.
See the Intake prototype →Nine hard flags are designed to run first as pre-condition checks, surfacing blockers before any classification is proposed. Weighted and readiness scoring then produce one PolarAIs Classification, applied the same way every time, so the plan holds up with the board and never depends on who is in the room.
See the Decision Card prototype →A deterministic rules engine is designed to read each entity's flags and scores and scope exactly the workstreams it needs, so no one starts from a blank template. AI tailors the task language to the entity and calibrates priority by revenue band and flag interaction, so a high-revenue seller with a banking blocker gets treasury and contracting elevated automatically.
See the Blueprint prototype →Every task is tracked to completion and hard-flag dependencies gate closure, so nothing closes out of order. Every decision is logged, timestamped, attributed, and immutable, so a program can change hands in month eighteen without losing a single decision.
See the Governance prototype →The Barycenter Engine is designed to hold your target end-state and the dollars to reach it, model the optimal structure across eight dimensions, and generate scenarios for speed, tax, or revenue. It rebalances the target as new facts arrive, so you always know where the portfolio is heading and what the gap is worth.
See the End-State prototype →“Optimal” isn't universal. The engine is designed to apply the acquirer's own priorities across 8 dimensions, legal, tax, GTM, certifications, vendor, compliance, ERP, and employee footprint, and generate three scenarios.
Wind down the simplest entities first to build momentum; accept some tax suboptimality for 6–12 months of program time.
Output: target entity count achievable within 18 months, quarter by quarter.
Identify the optimal IP holdco and design intercompany flows before any dissolution begins.
Output: recommended IP holdco jurisdiction and projected effective tax rate vs. current structure.
Identify the fewest Contracting Centers covering all required jurisdictions; clear revenue-entity flags first.
Output: Contracting Center count, novation sequence, and revenue % in optimized structure at 12/18/24 months.
Illustrative example, not a real customer or live platform data.
The Vantage Group is a hypothetical PE-backed platform company. It has just closed the acquisition of a mid-market software business that came with four legal entities attached: a US operating company that holds the customer contracts, a UK distribution subsidiary, an IP holding company, and a dormant shell left over from a prior owner’s own acquisition. Running that entity data through the Barycenter Engine, PolarAIs' end-state intelligence engine under a GTM/Revenue optimization objective produces the target-state recommendation below.
Each dispositioned entity generates its own set of owned, dated Layer 4 tasks directly from this table. Vantage US OpCo’s path to becoming the surviving Contracting Center is a 6-month workstream that unlocks an estimated $4.2M in cross-sell revenue previously blocked by the fragmented contracting structure. Torque Dormant Co.’s wind-down is a 12-month workstream that removes $95K a year in carrying overhead the acquirer inherited but never actioned.
Illustrative example, not a real customer or live platform data.
US — Technology — Stock Purchase
Illustrative example, not a real customer or live platform data.
Extending the Vantage Group scenario to its full post-close footprint: the platform inherited 47 entities across the current acquisition and two prior bolt-on deals it never fully integrated. The Portfolio Command View surfaces the whole picture at once.
Illustrative example, not a real customer or live platform data.
Entity intake generates structured fingerprints; diligence findings load directly into scored profiles before Day 1.
Complete profiles, run scoring, assess hard flags, assign classifications, design the full workplan before execution begins.
Execution proceeds against a documented plan. Hard flag dependencies gate closure. Every decision is logged.
Entities dissolve, merge, or stabilize per classification. Each completed integration compounds institutional knowledge.
Systems built to hold what a team or organization knows and put it back to use break into four working parts. Most get one or two right and treat the rest as someone's job to remember.
Getting signals, capturing what happens as it happens. PolarAIs: event capture from the platform's own task execution, plus scoped mailbox ingestion where a customer connects one.
Remembering, storing it durably and structured. PolarAIs: a seventeen-table ontology, meta- and instance-level, isolated per customer.
Dreaming and pruning, correcting, consolidating, and discarding what's gone stale. PolarAIs: a drift-detection job checking stored values against source systems or last-updated timestamps, on a cadence matched to how fast each field changes, flagged to a person, never auto-corrected.
Speaking and searching, surfacing the right piece at the right moment. PolarAIs: the Portfolio Command View and the scoring layer, both computed from current state on every read, never a stale cached number.
PolarAIs is designed to hold all four from the first customer, not to bolt the missing ones on after finding out the hard way what happens without them.
Talk through your entity footprint and see how the Barycenter Engine would score it.
Schedule a Consultation