Trust is not a badge. It is the decision record.
Fynally turns every travel request into a reconstructable approval record: source evidence, policy version, price snapshot, budget context, approver action, timestamp, and exportable audit trail.
- Same input + same policy = same outcome
- AI extracts; rules evaluate; approvers decide
- Every approval has evidence attached
Trip Approval Request
Trip
New York → London
14–18 Sep 2026
Airports: JFK → LHR
Traveler
Maya Chen
Trip reason
Customer Onboarding
Request cost
USD 1,284.20
Price breakdown
Fare: USD 731.00
Taxes / fees / carrier charges: USD 553.20
Benchmark price
USD 1,230.00
Policy
Exception acknowledged
Reason
Advance purchase window missed
Budget
CS-402 · Enterprise Success
USD 1,284.20 projected impact
Approval appr_2026_0448 · Policy policy_2026_05_03
Snapshot snap_8f42c1 · Evidence artifact_91d7a2
Approval ID: appr_2026_0448
Policy version: policy_2026_05_03
Price snapshot: snap_8f42c1
Source evidence: artifact_91d7a2
Requester: Alex Rivera
Approver: Dana Ross
Cost center: CS-402 · Enterprise Success
Budget owner: Dana Ross
Replay pointer: replay_2026_0142
Decision timestamp: 2026-06-18 14:32 UTC
Air price evidence
Fare USD 731.00 · Taxes/fees USD 553.20 · Benchmark USD 1,230.00
Hotel price evidence
Original quoted total: USD 1,256.27
Discount applied: USD 138.05
Room charges: USD 997.04
VAT: USD 101.66
Local tax / fees: USD 19.52
Request cost: USD 1,118.22
Every decision carries its own evidence.
Fynally does not ask Finance to trust a chat message, a forwarded email, or a manager’s memory. Each approval record is tied to the evidence that produced it.
- 01
Source evidence
Original email, PDF, screenshot, upload, calendar object, or Slack/Teams artifact linked to the intake record.
- 02
Policy version
The policy rules used at evaluation time are pinned, versioned, and reproducible.
- 03
Price snapshot
Benchmark context is captured with source, timestamp, request parameters, and returned price facts.
- 04
Budget context
Trip purpose, cost center, budget owner, budget period, and budget impact are captured as part of the finance decision record.
- 05
Decision payload
Approval, rejection, required changes, violations, rationale, approver identity, and timestamp.
- 06
Exportable audit trail
Finance can export decision records and evidence references for audits, due diligence, or month-end review.
No AI makes approval decisions.
Fynally uses AI only where it belongs: extracting and normalizing messy travel artifacts. Policy evaluation and enforcement are deterministic. Approval and rejection actions are governed, attributable, and evidence-backed.
AI may help with
- Reading booking confirmations
- Normalizing routes, dates, passengers, and prices
- Extracting trip purpose and request details from messy artifacts
- Explaining extracted details
- Flagging low-confidence fields for bounded clarification or governed exception routing
AI may not do
- Approve trips
- Reject trips
- Enforce policy
- Override Finance controls
- Create hidden decision paths
- Bypass required evidence
Why it matters
- Approvals must be accountable
- Rejections must be explainable
- Policy evaluation must be reproducible
- Finance needs evidence, not guesswork
- Auditors need a decision record, not an AI opinion
Decisions Follow Consistent Rules. Every Time. Explainable. Auditable.
The governed path from request to audit.
Fynally creates a controlled path from fragmented travel intake to finance-grade approval evidence.
- 1
Booking evidence received
Email, PDF, screenshot, upload, calendar object, or Slack/Teams artifact.
- 2
Trip candidate created
Incomplete or recoverable inputs create a candidate before becoming a canonical Trip.
- 3
Completeness checked
Missing or ambiguous fields trigger bounded clarification or governed exception routing, not silent approval.
- 4
Policy evaluated
Versioned, deterministic policy rules produce reason-coded outcomes.
- 5
Price snapshot captured
Benchmark evidence is stored with source and timestamp. If unavailable, the unavailable state is recorded with reason and audit context.
- 6
Approval card delivered
Slack/Teams card shows itinerary, trip purpose, cost context, policy result, violations, price delta, and required action.
- 7
Decision locked
Approval or rejection and the evidence record are preserved for export and replay.
Uncertainty does not pass silently.
Travel artifacts are messy. Screenshots are cropped. PDFs omit taxes. Email threads contain conflicting details. Fynally does not hide uncertainty or silently push questionable requests forward.
- Low-confidence or incomplete fields route to bounded clarification when recoverable.
- Unrecoverable, conflicting, unsupported, or unmapped cases become governed exceptions with traceable status.
- Missing required fields trigger clarification before approval handoff.
- Unmapped users create operator-visible exceptions.
- Rejections and escalations are traceable outcomes.
- Canonical Trips are created only after evidence is sufficient.
Security controls built for finance workflows.
Security is not treated as a separate promise. It is embedded into how travel evidence, approvals, and exports are controlled.
Tenant isolation
Company data is separated through multi-tenant controls and access boundaries.
Role-based access
Finance, approver, admin, and operator actions follow least-privilege access.
Minimal channel exposure
Slack/Teams cards show only what approvers need to decide. Sensitive evidence remains controlled.
Audit logging
Access, evaluation, approval, rejection, and export activities are logged for traceability.
Data minimization
Fynally stores what is needed to evaluate, approve, audit, and export travel decisions.
Production discipline
Controls are designed for monitored, governed, production workflows — not experimental approval behavior.
Audit packs without archaeology.
When Finance, Procurement, auditors, or diligence teams ask what happened, the answer should not require four people searching Slack, email, expense tools, and spreadsheets.
Example audit pack
Finance export-ready decision record
Maya Chen · JFK → LHR · 14–18 Sep 2026 · Approved by Dana Ross · USD 1,284.20
Request
Trip ID: trip_2026_0142
Approval ID: appr_2026_0448
Traveler: Maya Chen
Requester: Alex Rivera
Trip reason: Customer Onboarding
Route / dates: JFK → LHR · 14–18 Sep 2026
Finance controls
Cost center: CS-402 · Enterprise Success
Budget owner: Dana Ross
Projected budget impact: USD 1,284.20 projected impact
Decision record
Approval decision: Approved
Accountable approver: Dana Ross
Decision rationale: Client-facing implementation date fixed
Decision timestamp: 2026-06-18 14:32 UTC
Policy and price evidence
Submitted amount: USD 1,284.20
Benchmark price: USD 1,230.00
Policy result: Exception acknowledged
Policy violations: Advance purchase window missed
Policy version: policy_2026_05_03
Price snapshot: snap_8f42c1
Audit references
Source evidence: artifact_91d7a2
Replay pointer: replay_2026_0142
Approval ID: appr_2026_0448
Built as a governance layer, not another travel stack.
Fynally protects trust by staying focused. It governs travel approval decisions above existing booking, expense, TMC, and finance systems.
- Not a booking engine
- Not a ticketing platform
- Not a concierge desk
- Not an expense UI
- Not an email-first approval workflow
- Not a custom workflow engine
Focused scope makes the approval record cleaner, the controls stronger, and the finance evidence easier to defend.
Trust FAQ
No. AI may extract and normalize booking evidence. Rules evaluate. Accountable approvers decide.
It routes the record to bounded clarification when the issue is recoverable. Unrecoverable, conflicting, unsupported, or unmapped cases become traceable governed exceptions. It does not silently approve uncertain records.
Fynally records a governed unavailable state with reason, source, timestamp, and audit context. It does not silently omit price evidence.
Yes. Each decision is designed to link source evidence, policy version, price snapshot, budget context, approver action, rationale, and timestamp.
No. Fynally sits above booking and expense tools as a governance and evidence layer.
Slack and Teams are the governed approval and outcome surfaces.
Fynally stores what is needed to evaluate, approve, audit, and export travel decisions.
Give Finance a decision record it can defend.
Fynally is being built for companies that need travel approvals to be fast, controlled, explainable, and audit-ready from go-live.