Manual note-writing time cut to a fraction
Note drafting and transaction sorting automated on Bedrock
Planner review built into every AI output
Fearless Finance
Washington, D.C.
Financial Planning
Amazon Bedrock, Amazon Bedrock Knowledge Bases, AWS Lambda, Amazon API Gateway, Amazon DynamoDB, Amazon S3 (SSE-KMS), Amazon Cognito
Fearless Finance, a fee-only financial planning firm, engaged Avahi to automate two of its most time-consuming manual workflows: post-meeting note writing and bank transaction categorization. Every client session previously required a planner to spend 20 to 40 minutes drafting a summary note by hand, then manually sort transaction data before cash flow analysis could begin. Avahi designed and built an AI-powered solution on Amazon Bedrock that drafts structured summary notes from planner session input and automatically classifies bank transactions, with a planner review step built into both workflows. The result is a faster, more consistent documentation process that frees planners to spend more time with clients rather than on paperwork.
Fearless Finance is a fee-only, hourly financial planning firm based in Washington, D.C., regulated by the Securities and Exchange Commission. The firm serves thousands of client households through a team of financial planners who deliver comprehensive planning services, covering balance sheets, cash flow, retirement, insurance, estate planning, and investment guidance, entirely through virtual meetings.
As Fearless Finance grew, the administrative burden of post-meeting documentation became the primary constraint on planner capacity. After each client session, a planner would manually draft a summary note following the firm’s standardized template, a process that took 20 to 40 minutes depending on the complexity of the meeting. For cash flow review sessions, planners also had to manually sort each imported bank transaction into the firm’s expense categories before any analysis could begin.
Left unaddressed, this bottleneck would have limited how many clients each planner could serve as the firm scaled toward consolidating its planning team onto a single platform. Fearless Finance needed a way to reduce the time spent on documentation without compromising the quality or consistency of client-facing deliverables, all while meeting SEC requirements for data handling and record-keeping.
Because the engagement involves personal financial data under SEC oversight, Fearless Finance needed a platform that could meet strict encryption, access control, and record-keeping requirements. AWS provided the managed services to satisfy those requirements without added operational overhead: Amazon S3 with SSE-KMS encryption and timestamped storage for SEC record-keeping, TLS for all data in transit, and Amazon Cognito to control access to planner and client data.
Fearless Finance’s own development team was focused on building out the firm’s core planning platform and did not have the bandwidth or AI expertise to take on a generative AI build in parallel. Avahi’s experience designing and deploying Amazon Bedrock solutions, combined with a clear plan for integrating with the firm’s existing platform through well-defined APIs, gave Fearless Finance a low-risk path to add AI capabilities without diverting its own engineering team from platform development.
Avahi built two AI-powered capabilities on AWS, each with a human review step so planners stay in control of what gets sent to clients.
The note synthesis engine uses an Amazon Bedrock Knowledge Base, built from the firm’s meeting templates and planning methodology, to generate a structured draft summary note from planner-entered session data. Planners enter notes through a dedicated input form, review and edit the AI-generated draft in a purpose-built interface, and export a timestamped PDF for the client. Finalized notes are archived in Amazon S3 with SSE-KMS encryption and indexed for retrieval to support SEC record-keeping.
The transaction categorization tool processes client-uploaded bank transaction files and classifies each transaction into the firm’s expense categories using a three-stage approach: a local lookup against previously reviewed transactions, an Amazon Bedrock model for new transactions, and a third-party places lookup as a fallback for ambiguous merchants. Categorized results are presented back to the planner for confirmation or correction, and every correction feeds back into the local lookup so the system improves with use.
Both workflows run on AWS Lambda and Amazon API Gateway, with Amazon DynamoDB storing transaction and note metadata. The system was built to expose clean API endpoints so Fearless Finance’s own development team could integrate the AI services directly into the firm’s existing planning platform.
AI-powered note synthesis engine built on an Amazon Bedrock Knowledge Base
Planner note input form and a review and edit interface for AI-generated notes
Timestamped PDF export with SSE-KMS encrypted archival in Amazon S3
Bank transaction categorization tool using a three-stage classification pipeline
Planner-facing review interface for confirming or correcting categorized transactions
REST API layer for integration with the firm’s existing planning platform
Amazon Cognito authentication across both workflows
Final architecture documentation, deployment runbook, and source code handover
Fearless Finance planners tested the platform against real client sessions before sign-off, using it in parallel with the firm’s existing process. Planners found that once session data was loaded into the platform, generating and reviewing a note took a fraction of the 20 to 40 minutes the manual process required, even accounting for the review and edit step every note still goes through. The firm’s planners and leadership signed off on the solution after hands-on testing with live client data, and Fearless Finance is now evaluating an expanded, production-grade build to carry the capability forward.
Fearless Finance
Washington, D.C.
Financial Planning
Amazon Bedrock, Amazon Bedrock Knowledge Bases, AWS Lambda, Amazon API Gateway, Amazon DynamoDB, Amazon S3 (SSE-KMS), Amazon Cognito
Fearless Finance, a fee-only financial planning firm, engaged Avahi to automate two of its most time-consuming manual workflows: post-meeting note writing and bank transaction categorization. Every client session previously required a planner to spend 20 to 40 minutes drafting a summary note by hand, then manually sort transaction data before cash flow analysis could begin. Avahi designed and built an AI-powered solution on Amazon Bedrock that drafts structured summary notes from planner session input and automatically classifies bank transactions, with a planner review step built into both workflows. The result is a faster, more consistent documentation process that frees planners to spend more time with clients rather than on paperwork.
Fearless Finance is a fee-only, hourly financial planning firm based in Washington, D.C., regulated by the Securities and Exchange Commission. The firm serves thousands of client households through a team of financial planners who deliver comprehensive planning services, covering balance sheets, cash flow, retirement, insurance, estate planning, and investment guidance, entirely through virtual meetings.
As Fearless Finance grew, the administrative burden of post-meeting documentation became the primary constraint on planner capacity. After each client session, a planner would manually draft a summary note following the firm’s standardized template, a process that took 20 to 40 minutes depending on the complexity of the meeting. For cash flow review sessions, planners also had to manually sort each imported bank transaction into the firm’s expense categories before any analysis could begin.
Left unaddressed, this bottleneck would have limited how many clients each planner could serve as the firm scaled toward consolidating its planning team onto a single platform. Fearless Finance needed a way to reduce the time spent on documentation without compromising the quality or consistency of client-facing deliverables, all while meeting SEC requirements for data handling and record-keeping.
Because the engagement involves personal financial data under SEC oversight, Fearless Finance needed a platform that could meet strict encryption, access control, and record-keeping requirements. AWS provided the managed services to satisfy those requirements without added operational overhead: Amazon S3 with SSE-KMS encryption and timestamped storage for SEC record-keeping, TLS for all data in transit, and Amazon Cognito to control access to planner and client data.
Fearless Finance’s own development team was focused on building out the firm’s core planning platform and did not have the bandwidth or AI expertise to take on a generative AI build in parallel. Avahi’s experience designing and deploying Amazon Bedrock solutions, combined with a clear plan for integrating with the firm’s existing platform through well-defined APIs, gave Fearless Finance a low-risk path to add AI capabilities without diverting its own engineering team from platform development.
Avahi built two AI-powered capabilities on AWS, each with a human review step so planners stay in control of what gets sent to clients.
The note synthesis engine uses an Amazon Bedrock Knowledge Base, built from the firm’s meeting templates and planning methodology, to generate a structured draft summary note from planner-entered session data. Planners enter notes through a dedicated input form, review and edit the AI-generated draft in a purpose-built interface, and export a timestamped PDF for the client. Finalized notes are archived in Amazon S3 with SSE-KMS encryption and indexed for retrieval to support SEC record-keeping.
The transaction categorization tool processes client-uploaded bank transaction files and classifies each transaction into the firm’s expense categories using a three-stage approach: a local lookup against previously reviewed transactions, an Amazon Bedrock model for new transactions, and a third-party places lookup as a fallback for ambiguous merchants. Categorized results are presented back to the planner for confirmation or correction, and every correction feeds back into the local lookup so the system improves with use.
Both workflows run on AWS Lambda and Amazon API Gateway, with Amazon DynamoDB storing transaction and note metadata. The system was built to expose clean API endpoints so Fearless Finance’s own development team could integrate the AI services directly into the firm’s existing planning platform.
AI-powered note synthesis engine built on an Amazon Bedrock Knowledge Base
Planner note input form and a review and edit interface for AI-generated notes
Timestamped PDF export with SSE-KMS encrypted archival in Amazon S3
Bank transaction categorization tool using a three-stage classification pipeline
Planner-facing review interface for confirming or correcting categorized transactions
REST API layer for integration with the firm’s existing planning platform
Amazon Cognito authentication across both workflows
Final architecture documentation, deployment runbook, and source code handover
Fearless Finance planners tested the platform against real client sessions before sign-off, using it in parallel with the firm’s existing process. Planners found that once session data was loaded into the platform, generating and reviewing a note took a fraction of the 20 to 40 minutes the manual process required, even accounting for the review and edit step every note still goes through. The firm’s planners and leadership signed off on the solution after hands-on testing with live client data, and Fearless Finance is now evaluating an expanded, production-grade build to carry the capability forward.
Let’s explore your high-impact AI opportunities together in a complimentary session
