Multi-entity spend management must preserve which legal entity requested, authorized, paid, reimbursed, and posted each expense. Payhawk is on this list because it positions itself for multi-entity and international teams. Airbase is here because it puts requests, cards, reimbursements, and payables behind one approval chain, which matters when the entity is decided before money moves. Ramp is here because many multi-entity groups are really several US entities, and its published scope is US finance teams running cards, expenses, and payables together. Consolidated visibility must not erase local ownership.
Buyer scenario: payment and project entities differ
Consider an employee of one entity buying for a project owned by another, in foreign currency, through a card issued by the employing entity. A refund arrives after one entity closes and an ERP mapping fails only for another entity.
Triggers and wrong picks by entity structure
Payhawk. Trigger: the route matrix below shows a foreign subsidiary, a card that settles outside USD, or a local reimbursement rail in use every month, not once a year. Check: confirm each required country and currency against current Payhawk documentation and its connector to your ERP before the demo, not after. Wrong pick: a purely domestic group in one currency, where entity and cross-border depth adds configuration without fixing anything.
Airbase. Trigger: entities share suppliers and offices, and disputes arise over which entity should have bought something. Check: raise a request on behalf of a sister entity's project and confirm the entity chosen at request carries through to card or bill payment and to posting. Wrong pick: a group whose spend is mostly small out-of-pocket claims; pre-approval then slows employees without changing who owns the cost.
Ramp. Trigger: every entity is a US entity and the pressing need is cards and bill pay per entity. Check: confirm that each entity can be an applicant and how the ERP connection assigns the subsidiary on each charge. Wrong pick: a required non-US entity or currency route that current Ramp documentation does not confirm, or foreign rows that recur often enough that a manual exception would run as routine work rather than as an occasional, governed fallback. Ramp's published positioning centers on US finance teams, which is why an all-US matrix is this guide's preferred trigger; that positioning alone does not rule out a group with a few foreign rows.
Verify issuer eligibility, country support, and repayment terms separately for every entity.
Map employee, applicant entity, project entity, supplier, original currency, card, reimbursement, approval, and ledger company.
Build a route matrix before the demos. For every employee group, record the card applicant, settlement currency, reimbursement rail, approval chain, ERP company, and close calendar. Verify every required combination through current evidence. This prevents a group-level feature from receiving credit where the local path is unsupported or depends on an undocumented manual workaround.
Define group versus local permissions. Central finance may view consolidated spend but should not necessarily approve a local reimbursement or change a ledger mapping. Test whether the system can separate visibility, policy ownership, payment authority, and posting authority without blocking legitimate investigation across entities.
Add an inter-entity dispute to the trial. One entity may believe a purchase belongs to another after payment, while the card liability and employee relationship remain local. Record who may propose reassignment, who approves it, what evidence follows, and how ledgers receive the result. Software should preserve the original payment fact and the later allocation decision separately. A consolidated group view must not make local responsibility disappear.
Require every group report to drill back to the local source, original currency, payment rail, approval, and correction. Consolidation without that path can conceal offsetting errors between entities.
Decision criteria for entity ownership
Inspect entity onboarding, local card availability, limits, reimbursements, currency evidence, request routing, inter-entity coding, refunds, approvals, and close queues. Ask who can reassign an entity and how correction appears after approval.
For ERP connections, define local ledger, group dimensions, source of truth, latency, duplicate keys, rejection isolation, retry authority, and reconciliation. Group security features do not prove local compliance.
Reproducible multi-entity evaluation
Create users and approvers in separate entities. Request spend for another entity, pay by card, submit a personal reimbursement, split coding, change project ownership, process a foreign-currency refund, and close one entity first. Write expected local and consolidated states.
Trigger an invalid ERP mapping in one entity while another posts. Repair, replay once, and reconcile original currency, converted evidence, approval, refund, and ledgers. Spend Management Guide has not run this configured comparison.
Edge case: entity changes between authorization and settlement
Reassign the intended entity after card authorization but before settlement. Determine which entity owes the issuer, which receives the posting, and how intercompany treatment is exposed. Finance decides the accounting response.
Then add a new entity and verify that policies and dimensions require deliberate approval rather than invisible copying.
Verdict, and when the answer belongs in the ERP
Read the route matrix row by row. Non-US entity or non-USD rows that recur every month point to Payhawk. Rows that agree on currency but fight over which entity owns the purchase point to Airbase. Rows that are all US entities, USD, and card-funded point to Ramp, and Ramp stays first when the foreign rows are occasional, each required route is confirmed in current Ramp documentation, and a named payer, retained currency evidence, and a fixed posting rule cover the remainder. Whichever wins must keep applicant, payment, reimbursement, currency, and ledger ownership traceable through late events.
None of the three fits when the real problem is intercompany allocation, transfer pricing, or consolidation; those belong in the ERP and the group controller's process, and spend software should only feed them clean entity data. None fits either when a required country has no supported card or reimbursement rail from any candidate. Document that population's fallback instead of forcing a group-wide tool on it.
Traceable evidence
Sources for this decision
- vendorPayhawk official product sitePayhawk · checked Aug 5, 2026 · supports: Vendor-published product scope used to verify capabilities relevant to this buyer context: Multi-entity and international teams evaluating cards, expenses, and finance-system connections. It does not prove the guide's fit verdict, configured performance, current pricing or compliance.Open source ↗
- vendorAirbase official product siteAirbase · checked Aug 5, 2026 · supports: Vendor-published product scope used to verify capabilities relevant to this buyer context: Finance teams evaluating request-to-pay, cards, reimbursements, and accounts payable controls. It does not prove the guide's fit verdict, configured performance, current pricing or compliance.Open source ↗
- vendorRamp official product siteRamp · checked Aug 5, 2026 · supports: Vendor-published product scope used to verify capabilities relevant to this buyer context: US finance teams evaluating cards, expenses, controls, and accounts payable in one operating workflow. It does not prove the guide's fit verdict, configured performance, current pricing or compliance.Open source ↗