Assessment · 2 to 3 weeks
AI readiness assessment
Data foundation review, SAP landscape review, three candidate workflows, platform options and a cost model. Delivered as a written report and a board-ready briefing.
AI Agency
AI for the finance function, built to post.
Advisory and build work for organizations that want AI touching real documents, real bank statements and real approvals. We design the strategy, build the agents, and run our own products on the same stack, so we know exactly what holds up in production.
What we do
Advisory first. Then agents. Then products.
AI readiness and strategy
An honest assessment of your data foundation, your SAP landscape, and the first three workflows worth automating. Written for the CFO and CIO together, with a cost model and a governance plan.
SAP Joule, BTP and Business AI architecture
Where SAP's own AI belongs, where it does not, and how to place custom agents alongside it on BTP without breaking the clean core.
Agentic workflows for finance
AP invoice triage and coding, bank statement reconciliation, cash flow forecasting, month-end variance narratives, audit evidence retrieval. Deterministic where money moves, generative where judgment helps.
Retrieval over enterprise documents
Planner, retriever, reasoner and validator agents over contracts, policies, statements and spreadsheets. Grounded, cited answers and a grounded refusal when the source is not there.
Enterprise AI infrastructure and platform advisory
Platform selection across on-premises, cloud and hybrid: NVIDIA DGX, Dell AI Factory, HPE Private Cloud AI, Nutanix, VAST Data, IBM watsonx, and the hyperscalers. LLM strategy, RAG architecture, GPU sizing and cost modeling.
Custom AI product build
From prototype to launch: product definition, engine, interface, deployment and operations. We have shipped an options intelligence platform, an AI CFO agent and a content engine of our own.
MCP connectors and tool integration
Model Context Protocol servers that give agents safe, scoped access to SAP, mail, calendars, banks, market data and internal tools. We built a 78-tool MCP server for a trading platform; we can build one for yours.
AI governance and change management
Model risk, audit trails, separation of duties for agents, and the training that turns a pilot into a habit. The same discipline we bring to a SOX-scoped SAP program.
Founder and business operating systems
For smaller organizations: a Claude Code based business operating system with your context, connectors, skills and memory, so one person can run like a team.
How we build agents
Every action has a gate. Every answer has a source.
Finance does not forgive confident mistakes. Our agent designs borrow the controls finance already trusts: separation of duties, validation before posting, and an audit trail for every decision.
- Deterministic rules wherever money moves; the model proposes, the engine decides
- A validator agent checks every output against the source before it is shown
- Human approval gates at the points an auditor would expect them
- Every recommendation shows what was considered and rejected, and why
- Learned corrections: a decision is made once and never asked again
Flagship advisory · April 2026, version 2.0
Enterprise AI Infrastructure and Platform Advisory
A complete landscape assessment for C-suite and IT leadership: every major on-premises, cloud and hybrid platform, LLM strategy, RAG architecture, GPU infrastructure and deployment models, with cost modeling for Fortune 500 and mid-market profiles.
- Adopt a hybrid, multi-model architecture as the default; route workloads by data sovereignty, volume and latency
- Select platforms on your existing ecosystem, team and data scale, not vendor marketing
- Invest in the data layer first and the compute layer second; that is where AI deployments actually fail
How we engage
Three ways in.
Sprint · 4 to 8 weeks
Agentic workflow sprint
One workflow to production, with validators, gates, runbooks and training. Fixed scope, fixed team, weekly demos on your data.
Retained · Monthly
Fractional AI lead
A senior AI transformation lead embedded with your finance and IT teams: roadmap ownership, vendor management, architecture decisions and governance.
Next step
Which workflow should go first?
Bring the process that costs your team the most hours each month. In 30 minutes we will tell you whether an agent can own it, what it needs from SAP, and what it would take to ship.