By James Aspinwall — February 2026
Solaris (formerly Solarisbank) is a fully licensed German bank offering Banking-as-a-Service. They provide the regulated infrastructure — accounts, cards, lending, KYC — so that fintechs and non-financial companies can embed banking into their own products via API. Think of them as the engine under someone else’s hood.
They hold a full German banking license and operate under BaFin supervision, which means every engineering decision carries regulatory weight.
Vendor Relationships
Solaris shows up as a customer or case study across several major platforms:
- Atlassian — featured on their customers page
- AWS — published case study, “Germany’s First Bank in the Cloud”
- Slack — customer story on partner communication via Slack Connect
- IDnow — case study on digital onboarding and video identification
- Netguru — client reference for backend engineering on debit cards and loans
- Mambu — core banking platform partner, powers their lending and BNPL subledger
Tech Stack
From StackShare and the Solaris engineering blog:
- Languages: Ruby, Elixir, Kotlin, Java, Python, JavaScript, TypeScript, Golang
- Infrastructure: AWS (including Lambda), Docker, PostgreSQL, MySQL
- Architecture: Transitioning from monolith to microservices
- Notable: The debit card platform runs on Ruby. The consumer loan engine runs on Elixir.
Why This Matters
Solaris uses Elixir in production for their lending products. That’s a direct conversation opener — WorkingAgents runs on Elixir, and Solaris already has engineers who understand the BEAM ecosystem. No technology evangelism required. The pitch lands on familiar ground.
The BaFin regulatory pressure is real. They’re a fully licensed bank running cloud-native infrastructure under heavy compliance requirements. That’s the profile that needs AI governance tooling but likely can’t justify building it internally — exactly where an integration consultancy adds value.
Consulting Positioning
The agents are not generic AI assistants. They are domain-specific adapters — “AI that speaks BaFin.” Each agent integrates with industry-standard submission systems (goAML, Bundesbank ExtraNet, COREP filing), uses regulatory templates (Verdachtsmeldung format, DORA incident reports), and enforces jurisdiction-specific rules (GwG thresholds, CRR limits, MaRisk escalation chains).
This is where generic AI falls flat. BaFin reporting, SOX compliance, HIPAA logs, financial audit trails — these require regulatory knowledge baked into the agent, not bolted on as a prompt. Each industry vertical becomes a specialized offering with pre-built integrations. Stickier, harder to replace, premium pricing justified.
The pitch to Solaris: “AI that not only tells you what to do, but can do it for you — with appropriate oversight.”
Deployment & Pricing Model
Each customer gets a dedicated instance — one Docker container in its own VPC. No shared infrastructure, no multi-tenant complexity.
Per-Instance Pricing
- Base instance — covers VPC, compute, database, and storage
- AI usage — metered based on API calls (LLM invocations per agent)
- Storage — tiered based on data volume (transaction history, audit trail retention)
- Support tiers — community, business, enterprise
Enterprise Tier
For organizations running multiple business units (or multiple regulatory jurisdictions):
- Multiple instances under a single contract
- Consolidated billing across all instances
- Shared support and success management
- Custom SLAs per instance
Infrastructure Cost Model
Transparent cost breakdown per instance. The customer sees exactly what they pay for — no opaque “per-seat” licensing, no hidden platform fees. Costs scale with usage (transaction volume, agent invocations, data retention), not headcount.
Volume discounts apply for customers running multiple instances. The marginal cost of an additional instance is infrastructure only — the agent configurations, system prompts, and regulatory integrations are the same image.
Missing Operational Data
We cannot build a credible cost-reduction model without Solaris-specific numbers:
- Daily operations: Infrastructure scale, incident volume, service scope, BaaS partner count, transaction throughput
- Staffing: Headcount per function — compliance analysts, risk officers, IT security, KYC reviewers
- Alert volume: AML alerts/month, API incidents/month, exposure limit approaches, KYC applications processed, regulatory changes assessed per year
- Cost structure: Operational headcount and rough salary bands for ROI calculation
The demo proves the technology. The business case requires their baseline. Without it, we are guessing at the value proposition.