Making US tax preparation less fragmented with AI
Tax practitioners spend hours collecting documents, reconciling financial information and reviewing data across disconnected tools. The brief was to build an AI-assisted workflow that brings these steps together, while keeping practitioners in control of every decision.
ROLE
Product Designer
COMPANY
SignalsHQ
TIMELINE
6 months
TOOLS
Figma, Claude Design & Code
About the Company
SignalsHQ is an early-stage, pre-seed startup building automation tool for US tax workflows. The workflows are data and client management, document verification and prepare return draft.
DATA & CLIENT MANAGEMENT
Collect client information, prior-year returns, and supporting documents.
EXTRACT & REVIEW
Classify documents, extract key data, and review information for accuracy.
TAX RETURN PREPARATION
Use verified client data to prepare and move forward with the tax return.
My Role
Discovery
Collaborated closely with the Product Manager to define user flows and end-to-end journeys that accurately reflected practitioner behavior and refined PRD.
Worked with an in-house Tax Practitioner to establish the technical structure and data schema underlying the product, drawing on their domain expertise and field-specific technicalities.
Development
Collaborated with developers to ensure smooth technical implementation, adjusting designs to accommodate tech constraints and limitations.
Collaborated with the Product Manager to define MVP scope and prioritize features for release.
Why worth solving?
Tax preparation looks straightforward on paper: collect documents, review information and file returns.
But conversations with practitioners revealed something different.
Practitioners move information from forms and bank statements into their workflows. This takes time away from higher-value review work.
Fragmented workflows
Client informations are spread across emails, spreadsheets, folders and tax software. Practitioners piece together the information they need.
I spoke with practitioners across different firm sizes to understand how they manage this process, where work gets stuck and what they need from an AI-assisted system.
How did we approach research?
We triangulated insights from multiple sources:
12 Solo CPAs and small firms | 9 Mid-sized
But, who are practitioners?
Tax practitioners are the primary operators of the tax filing workflow. They can be either certified or non-certified.
Key Findings
Three insights strongly shaped the product direction:
The prior-year return (PYR) is foundational
What we observed: Practitioners rely on prior-year returns to understand client information and identify what's changed.
Why it matters: Each stage should carry PYR information, overwrite or counter ask at times.
Workflows are asynchronous
What we observed: Documents arrive at different times, and practitioners revisit earlier work as new information comes in
Why it matters: Each stage should update whenever new data comes in.
App needs to leave evidence of work
What we observed: Practitioners trust on source of truth either standard doc or info shared on call.
Why it matters: Review workflows depend on auditable chains from source document to extracted value to output form.
Problem Statement
Tax preparation is rarely a linear process.
Practitioners receive documents in pieces, reconcile information across spreadsheets and tools, and repeatedly follow up with clients to fill in the gaps.
Existing tax software helps with filing, but much of the preparation work still happens manually.
The opportunity was to bring these fragmented workflows together, automate workflows without taking control away from the practitioner.
JBTD
Finalizing scope
Based on practitioner interviews, workflow observations and Jobs to Be Done, we defined success around making tax preparation less manual, more structured and easier to validate.
Reduce client intake effort: Decrease the time practitioners spend creating document requests, collecting information and following up on missing items.
Reduce manual data processing: Decrease the time spent classifying, extracting and organising information from standard and non-standard tax documents.
Accelerate tax return review: Reduce the time required to consolidate current-year financial information into a structured, review-ready workspace.
Improve missing-data and discrepancy detection: Increase the number of missing documents, incomplete fields and year-over-year variations identified before review.
Interface Design
Designing the experience
I designer interface to manage Client document requests, aimed at improving efficiency by providing status, updates and actions to do next.
Due to NDA I can only show few information, for detailed case study, please reach out to me on
Outcome
All's well that ends well?
57%
24-Feedback responses
Across the three workflow stages
3-Areas assessed
Collection, review and preparation
Learnings
Task flow related interview turned out well but we lacked field-specific technicalities which was necessary of us to build few features as a team (backend & design). Deep dive into market research helped us getting into field domain and then ask behavioral questions.
Thank You
Thanks for Visiting!