'%20/%3e%3c/svg%3e){kind=small}
Preview
LLM router
24/7
The AI engine behind every automated decision
Payouts.com runs its AI on infrastructure built for reliability, control, and cost transparency — so the fraud checks, tax reviews, invoice reading, and follow-up that power the platform keep working even when a model provider doesn't.
AI Infrastructure
Reliability through routing
The LLM router provides automatic failover across providers — Vertex AI, Anthropic, Gemini, Grok/xAI, OpenAI-compatible endpoints, and OpenRouter — so a single provider outage doesn't stop your finance operations.
Automatic provider failover
The LLM router fails over across Vertex AI, Anthropic, Gemini, Grok/xAI, OpenAI-compatible endpoints, and OpenRouter, so a single provider outage never stops finance operations.
Agent roster
Working
R
Running
ReconcilerMatching 84 invoices · 3-way
P
Running
PayerReleasing batch · 248 vendors
C
Running
ChaserChasing 12 approvals · WhatsApp
Tasks completed today12,840
Re-point models without code
The LLM Engine console lets staff re-point any AI feature to a different model without changing code — no redeploy required.
Policy guardrail
Preview
Pay · Cedar Foods LLC
$4,200
$0Limit $10,000
Auto-approved · within policy
Auto-approved within policy98.2%
Every call site, controllable
A call-site inventory registers every AI usage point, so each one can be governed and controlled individually.
24/7 activity
Uptime 99.99%
00:0006:0012:0018:0024:00
Actions / min1,240
Cost & credits
Cost you can see and shape
Cost & usage observability tracks per-call latency, tokens, and cost. A credits economy attributes AI spend per client, and a cost simulator / optimizer models and reduces AI costs before they're incurred.
Orchestrator
dispatching
AP
AR
Onboard
Approve
Payout
Match
What you get
Smarter retrieval, grounded answers.
A semantic engine powers semantic search and matching across documents and vendors, and web grounding lets the AI ground answers in live web data where allowed.
Cost & usage observability
Track per-call latency, tokens, and cost across every AI feature.
Credits economy
Attribute AI spend per client with a built-in credits economy.
Throughput · preview
tasks / hour
0tasks / hr now
Hours saved this week1,280
AI engine at a glance
Your AI, routed and observed
24/7always available
80%of AI cost tracked
5+model providers
100%calls observable
'%20d='m0%200%2018%2011h65%2065%2065%2065%2066L51%2039h65%2065%2065%2065L18%2066h65%2065%2065%2065%2066L51%2094h65%2065%2065%2065L18%20121h65%2065%2065%2065%2066L51%20149h65%2065%2065%2065L18%20177h65%2065%2065%2065%2066L51%20205h65%2065%2065%2065L18%20232h65%2065%2065%2065%2066z'/%3e%3c/svg%3e)
'%20transform='translate(-20)scale(.72249)'%3e%3cpath%20fill='%23fff'%20d='M0%200h1063v708.7H0z'/%3e%3cpath%20fill='%23df0000'%20d='M0%200h1063v354.3H0z'/%3e%3cpath%20fill='%23fff'%20d='M245.2%2059.4a124.6%20124.6%200%200%200%201.1%20243.9%20126.9%20126.9%200%201%201-1.1-243.9'/%3e%3cpath%20fill='%23fff'%20d='m202%20162.4-18.9-13.8%2023.5-.2%207.2-22.3%207.5%2022.3h23.4l-18.8%2014%207.2%2022.3L214%20171l-19%2013.8zm26%2076.9-19-13.8%2023.5-.2%207.3-22.3%207.4%2022.2h23.5l-19%2014%207.3%2022.3-19-13.6-19%2013.8zm86.3-.6-19-13.8%2023.4-.2%207.3-22.3%207.4%2022.3H357l-18.9%2014%207.3%2022.3-19.1-13.7-19%2013.8zm25.7-76.2-19-13.8%2023.5-.2%207.2-22.3%207.5%2022.2h23.4l-18.8%2014%207.2%2022.3-19.1-13.6-19%2013.8zM271.7%20112l-19-13.8%2023.5-.2%207.3-22.3%207.4%2022.3h23.5l-19%2014%207.3%2022.2-19-13.6-19%2013.8z'/%3e%3c/g%3e%3c/svg%3e)
'%20transform='translate(-128)scale(.72249)'%3e%3cpath%20fill='%23fff'%20d='M0%200h1063v708.7H0z'/%3e%3ccircle%20cx='523.1'%20cy='344.1'%20r='194.9'%20fill='%23bc002d'%20transform='translate(-59.7%20-34.5)scale(1.1302)'/%3e%3c/g%3e%3c/svg%3e)
+33 more
Cost simulator & optimizer
Model and reduce AI costs before they're incurred.
Semantic engine & web grounding
Semantic search across documents and vendors, with answers grounded in live web data where allowed.
One AI engine
One engine behind every AI decision
Every fraud check, tax review, invoice read, and follow-up runs through one AI engine — routed across providers with automatic failover, observed per call for cost and latency, and re-pointable to a different model without code.
Shared task queue · auto-dispatch
0 queued
- Reconcile AR ledger · USD
- Chase 8 overdue invoices
- Classify 142 transactions
- Onboard vendor · KYB
- Release SEPA batch · €1.2M
- Match 84 invoices · 3-way
- Post GL entries · period
- Flag 3 duplicate bills
- Agent · AP Waiting for work Idle
- Agent · AR Waiting for work Idle
- Agent · Ops Waiting for work Idle
48,200tasks automated
9,120hours saved
24/7always running
FAQ
Frequently asked questions
Everything you need to know about the LLM router, AI cost and credits, and the controls behind every automated decision.
Still have questions?
What happens if an AI provider goes down?
The LLM router automatically fails over across multiple providers, so the platform's AI features keep operating without a single point of failure.
Can the model behind a feature be changed?
Yes. Staff can re-point any AI feature to a different model from the LLM Engine console without code changes, and every AI call site is individually controllable.
Is AI cost visible per client?
Yes. Per-call cost, latency, and tokens are tracked, and a credits economy attributes spend to each client.
Which model providers does the router support?
The LLM router works across Vertex AI, Anthropic, Gemini, Grok/xAI, OpenAI-compatible endpoints, and OpenRouter, so AI features keep running if any one provider fails.
Can the AI search documents and use live web data?
Yes. A semantic engine powers semantic search and matching across documents and vendors, and web grounding lets the AI ground answers in live web data where allowed.
A reliable AI engine
behind your finance ops
A multi-provider LLM router with failover, per-call cost observability, and a credits economy — re-point any AI feature without code. Talk to our team.