Agentic AI for Financial & Business Planning Document Building

We are looking for an agentic AI engineer to build a custom agent system that produces client ready feasibility studies, business plans, and financial projections. What the system needs to do: We will provide the templates, multipage prompts, CSV data files, and a client intake form. The agent needs to take those inputs and generate each document in our exact format and structure. Output has to be accurate and high fidelity, not generic filler since this will be used with actual lenders and clients to speed up our workflow. Every document must be specific to the client location we provide and dataset in the provided files. That means the agent performs real market research on that geography and pulls it into the analysis, not placeholder assumptions. Once the first draft is generated, it gets passed to a second agent. That agent runs a series of custom coded prompt steps that review, refine, and tighten the document. On average this loop should run about six passes before the final file is delivered. What we need from you: -Experience building multi-agent workflows, not single prompt wrappers -Ability to work from our existing templates and prompt libraries rather than replacing them. Our assumptions MUST be used to maintain accuracy and must NOT be disregarded. -Research and data retrieval integration for location-specific market analysis. We need up to date and accurate data going into these. -Structured document output that holds formatting across long files. Some of our files and business plans go up to 40-50 pages. -Financial modeling logic that produces consistent, defensible numbers that match the business plan and feasibility study. ALL numbers must be congruent and match up. -A refinement loop you can code, control, and adjust as we iterate. We are looking for ongoing updates and support to fine tune this as needed to ensure consistent and accurate results every time. -This MUST be easy for non-tech users to use and generate reports.