Introducing Holly
Instead is a tax-planning platform for accounting firms. As the company scaled, onboarding each client took 3–5 hours of manual data entry.
I led development of an AI-assisted system that reads prior-year tax returns, detects entities, and structures data automatically cutting setup time by 92% and enabling 400+ firms to onboard clients in minutes with full transparency and control.
Instead is a tax-planning platform for accounting firms. As the company scaled, onboarding each client took 3–5 hours of manual data entry.
I led development of an AI-assisted system that reads prior-year tax returns, detects entities, and structures data automatically cutting setup time by 92% and enabling 400+ firms to onboard clients in minutes with full transparency and control.
92% faster onboarding
from 3–5 hours to under 15 minutes.
Self-onboarding at 68%
onboard clients without customer support intervention.
7x more clients handled
during onboarding cycles.
76% template reuse rate
for saved templates, cutting repeat setups by hours.
94% fewer data-entry errors
with zero data loss across 400+ onboarded firms.
How it started
How it started
I shadowed a tax pro.
Three monitors open.
IRS portal. PDF return. Instead.
Every line she typed already existed.
I shadowed a tax pro.
Three monitors open.
IRS portal. PDF return. Instead.
Every line she typed already existed.
3–5 hours per client!
3–5 hours per client!
Manual setup. Repeated hundreds of times.
Manual setup. Repeated hundreds of times.


The Old Way
Each business entity - individual, corporation, partnership had to be manually entered.
No data recall.
No intelligence.
No automation.
Each business entity - individual, corporation, partnership had to be manually entered.
No data recall.
No intelligence.
No automation.
Onboarding was supposed to be simple. Until it started costing the company money.
Onboarding was supposed to be simple. Until it started costing the company money.
The Idea
The Idea
I believed onboarding didn’t need more forms - it needed memory.
What if we could read the returns itself, detect entities, and pre-fill everything?
One upload. Zero re-entry. Full control.
I believed onboarding didn’t need more forms - it needed memory.
What if we could read the returns itself, detect entities, and pre-fill everything?
One upload. Zero re-entry. Full control.


AI Document Drawer
AI shouldn’t feel like a feature - it should feel like memory.
The drawer became the quiet bridge between automation and trust: it reads, detects, and prepares everything, but never acts without user intent.
AI shouldn’t feel like a feature - it should feel like memory.
The drawer became the quiet bridge between automation and trust: it reads, detects, and prepares everything, but never acts without user intent.
Manual Setup
In fintech, control equals confidence.
The right side remains fully manual — not because AI couldn’t do it, but because users needed to see that they still could.
In fintech, control equals confidence.
The right side remains fully manual — not because AI couldn’t do it, but because users needed to see that they still could.
When Data Starts to Organize Itself
When Data Starts to Organize Itself
The entity view was designed for clarity at scale
Every client, entity, and sub-entity neatly structured for instant understanding.
It’s not just a list - it’s a map of relationships that grows as the business does.
The entity view was designed for clarity at scale
Every client, entity, and sub-entity neatly structured for instant understanding.
It’s not just a list - it’s a map of relationships that grows as the business does.
The entity view was designed for clarity at scale
Every client, entity, and sub-entity neatly structured for instant understanding.
It’s not just a list - it’s a map of relationships that grows as the business does.





The left sidebar isn’t static - it’s alive.
Each document uploaded reshapes the hierarchy in real time: new entities, updated savings, and refined structures - all of which can happen at any time, without breaking flow.
The left sidebar isn’t static - it’s alive.
Each document uploaded reshapes the hierarchy in real time: new entities, updated savings, and refined structures - all of which can happen at any time, without breaking flow.
The left sidebar isn’t static - it’s alive.
Each document uploaded reshapes the hierarchy in real time: new entities, updated savings, and refined structures - all of which can happen at any time, without breaking flow.
Designing for a Proactive AI
AI shouldn’t wait for instructions - it should anticipate them. I wanted the system to think before the user did.
In traditional tax workflows, professionals hunt through data, compare strategies, and calculate savings manually. Here, AI does that groundwork quietly in the background, pre-selecting the most relevant tax strategies based on uploaded documents.
AI shouldn’t wait for instructions - it should anticipate them. I wanted the system to think before the user did.
In traditional tax workflows, professionals hunt through data, compare strategies, and calculate savings manually. Here, AI does that groundwork quietly in the background, pre-selecting the most relevant tax strategies based on uploaded documents.



Instead of a long list of options, the interface opens with projected savings and pre-selected strategies, turning choice into confirmation.
Instead of a long list of options, the interface opens with projected savings and pre-selected strategies, turning choice into confirmation.
Built for Speed & Efficiency
Built for Speed & Efficiency


Direct Integration
Direct IRS and QuickBooks integrations pull existing client data instantly - no re-entry, no setup from scratch.


Save and Reuse
Any onboarding can be saved as a template and loaded again, keeping structure and settings consistent across clients.
The Result
92% faster onboarding
from 3–5 hours to under 15 minutes.
Self-onboarding at 68%
firms now onboard clients without customer support intervention.
7x more clients handled per tax professional
during onboarding cycles.
76% template reuse rate (as of November 2025)
for saved templates, cutting repeat setups by hours.
94% fewer data-entry errors
with zero data loss across 400+ onboarded firms.
92% faster onboarding
from 3–5 hours to under 15 minutes.
Self-onboarding at 68%
onboard clients without customer support intervention.
7x more clients handled
during onboarding cycles.
76% template reuse rate
for saved templates, cutting repeat setups by hours.
94% fewer data-entry errors
with zero data loss across 400+ onboarded firms.
92% faster onboarding
from 3–5 hours to under 15 minutes.
Self-onboarding at 68%
onboard clients without customer support intervention.
7x more clients handled
during onboarding cycles.
76% template reuse rate
for saved templates, cutting repeat setups by hours.
94% fewer data-entry errors
with zero data loss across 400+ onboarded firms.