AI Architecture · Multi-tenant · Staff IC

Designing Multi-Tenant AI for Professional Accountants

The challenge: an AI built for the wrong person

Intuit's GenAI framework was originally built for Small Business Owners — people who benefit from basic guidance like tax term explanations and simple categorization tips. When professional accountants, who manage hundreds of clients simultaneously, used this same AI, the results were jarring.

The AI couldn't distinguish between the accountant and their client. It gave CPAs advice meant for beginners. It had no concept of "I'm working across 200 client files right now." And it forced accountants to manually hop into each individual QuickBooks file — one at a time — to perform any real professional work.

I identified this gap, named it the "Identity Crisis," and argued to the business that we were actively alienating our most powerful users.

The framework: three architectural principles

Rather than designing surface-level UI fixes, I defined three foundational principles that would govern AI system behavior at the architecture level.

The delivery: three levels of execution

Level 2 — Iteration 1: Client dropdown in the input field

Shipping useful functionality while tech was developing

Engineering was actively building multi-tenant infrastructure, but it wouldn't be ready for months. I designed a pragmatic first iteration: a client dropdown selector built into the input field. Once selected, the client name converted to static text — making the scope explicit and immutable for that conversation.

Design Intuit Intelligence — Embedded in IAS Dashboard

The AI panel sits alongside the full Accountant Suite — giving accountants instant access to intelligence without breaking their workflow context. The "My practice" scope chip anchors the query to the right data environment.

Design Intuit Intelligence — Client context dropdown

The "My practice" chip expands to a searchable client list. Selecting a client scopes the AI query to that file — a pragmatic shipping path while multi-tenant infrastructure was being built.

Level 3 — Iteration 2: The "@" mention — the actual solve

Designing for how professionals actually think and work

The dropdown worked, but it felt clunky. I challenged the "one context at a time" limitation by designing the "@" mention feature — letting accountants reference specific clients inline within their query, just like tagging someone in Slack.

From the IAS dashboard, an accountant types "@Client A @Client B @Client C — compare Q3 revenue trends." The AI respects each client's permission boundary, synthesizes cross-client data, and returns an insight in seconds — a task that previously took 45–90 minutes of manual file navigation.

Design Intuit Intelligence — @ mention client picker

Typing "@" triggers a client picker dropdown. The accountant selects clients inline — each becoming a scoped reference in the query — enabling multi-client synthesis without leaving the IAS dashboard.

Concept Persona-aware conversation — the delivery

Same AI, two outcomes: without persona it talks like a beginner coach; with accountant persona + @ mentions it answers as a professional across the clients you tag.

Impact

The three architectural principles I defined became the governing framework for how Intuit AI behaves for all professional users — a reusable foundation that any future AI feature in the accountant ecosystem builds on.

Cross-client workflows that previously required hours of manual file navigation were reduced to seconds of natural language query.

What I learned

The hardest part of this project wasn't the design — it was the advocacy. Convincing a business that its most powerful AI feature was alienating its most valuable users required framing the problem in business terms: these accountants manage the majority of data in the Intuit ecosystem, and if they don't trust the AI, adoption fails.

Platform · 0→1 · Enterprise · Staff IC

Intuit Accountant Suite

The challenge: a strategic glass ceiling

QBOA had become the dominant accounting software for small firms — so dominant that the market was effectively saturated. There was nowhere left to grow in that segment. To generate meaningful new revenue, Intuit needed to move upstream into mid-to-large accounting firms: a completely different customer with far more complex needs. But the existing QBOA wasn't built for them. It couldn't run a firm. It couldn't manage teams, track service delivery across hundreds of clients, or give partners the visibility they needed to run a practice at scale. The gap was clear — but so was the size of the problem. Research showed these larger firms were stitching together ~11 different tools just to operate. For six months, leadership couldn't agree on how to move. The vision existed in 40-page strategy documents that no one could fully hold in their head at once. I decided to change the medium.

The intervention: prototype as executive alignment tool

I built a functional, high-fidelity coded prototype of the entire Intuit Accountant Suite — not a slideshow of screens, but a clickable, working system that let you navigate from client acquisition through service delivery to AI-powered insights in a single session.

The effect in the room was immediate. Seeing the 11-tool burden consolidated into one fluid experience moved the conversation from "should we do this?" to "how do we start?" Within weeks, the investment was approved and a new business unit was being staffed.

Screen 1Marketing & onboarding

End-to-end experience starting from the marketing page through to the live product.

Screen 2Sign-in & acquisition

Sign-in flow with client migration offer — reducing switching friction for large firms.

Screen 3HomeBase — firm dashboard

Priority tasks, calendar, disconnected accounts, and product updates in one view.

Screen 4Client AI profile

Profit trend chart + AI-generated advisory talking points per client.

Four screens from the functional coded prototype — built end-to-end to make the vision tangible for executive decision-makers.

The four pillars

1. Unified Client Management — end-to-end lifecycle

I designed a unified Client Lifecycle Engine that merged lead generation, onboarding, and active service delivery into a single persistent profile — giving partners a 360-degree view of client health, financial trends, and proactive next best actions.

Prototype Client list — unified roster

Every client in one place — accounting solutions, products, attached apps, and billing type visible at a glance. No CRM required.

Prototype HomeBase — firm-wide intelligence

Full firm health in one dashboard — P&L, books close, disconnected accounts, and work requests.

2. Integrated Service Management & Delivery

Accountants were jumping between management tools (Asana, Karbon) and accounting tools (QuickBooks) — two context switches per task, hundreds of times a day. I integrated service delivery directly into the management interface, so accountants could monitor their queue and execute tasks without leaving a single screen.

Prototype Books Close — Service delivery overview

The Close overview gives partners a firm-wide view of all client close engagements — period, due date, assigned team members, status, and task completion. Blocked clients surface immediately.

3. AI Data Intelligence (Client Console)

Mature firms sat on mountains of dormant client data with no way to use it strategically. I designed the Client Console to leverage AI for proactive Next Best Actions — elevating the accountant from reactive bookkeeper to high-value strategic advisor.

Prototype Client financial insights — portfolio table

300 clients queryable across expenses, revenue, current ratio, net profit margin, and AR past due — with colour-coded anomaly signals throughout.

Prototype Client P&L — AI anomaly drill-down

Clicking a flagged metric surfaces an AI insight card with trend history, root cause explanation, and a direct path to Ask Intuit Intelligence.

4. Team & Workload Management

Partners had no visibility into whether their team was over- or under-utilized. I designed a Team Management layer that lets partners track workload distribution, assign roles, and monitor staff productivity in real time.

Design provocations

Provocation Team management — roster & performance

Role, performance, hours, and AI-driven insights per staff member — a real-time view of who is doing what and where coaching is needed.

Provocation Capacity planning — weekly allocation

A timeline view of staff capacity across weekly periods — flagging time-off conflicts, over-allocation, and gaps before they become delivery problems.

Design leadership: co-developing with the market

I established a Co-Development Framework with leaders from mid-to-large accounting firms to ensure we were solving their most expensive problems before committing to production. I ran intensive workshops to map the 11-tool fragmentation in detail — those workshops directly produced the four pillar structure.

Impact

The prototype I built ended a six-month strategic stalemate and secured investment for a dedicated 100+ person organization. The suite launched and reached 10,000 firm adoptions within 4 months — positioning Intuit as a credible competitor in the large-firm market for the first time.

Analytics · AI-native · Data Platform

AI-Powered Portfolio Intelligence

Transformed QuickBooks from a siloed per-client based reporting system into a queryable, firm-wide intelligence platform — giving accounting firms their first portfolio-wide view across 250–1,800+ clients.

How it works

Accountants query any client data in natural language, filter by custom fields, and drill into AI-surfaced anomalies — all without leaving the Console or opening individual client files.

Future direction

The next generation moves from a queryable data table to a fully agentic practice management layer — proactively identifying what matters, ranking it by urgency, and generating analysis alongside the accountant in real time.

1 of 7

Console overview

2 of 7

Anomalies feed

3 of 7

All dashboards

4 of 7

High anomaly drill-down

5 of 7

Advisory Agent — prioritized list

6 of 7

Client analysis — inline expand

Future direction

Advisory Agent — Conversational dashboard construction

The accountant uploads a timesheet screenshot. The Advisory Agent reads the image, extracts metrics, proposes additional columns, and builds the dashboard schema — waiting for confirmation before populating.

←

→

Design Systems · Platform · Cross-ecosystem · Staff IC

A Unified Vision for Work Management

The challenge: the fragmentation tax

As Intuit grew its product portfolio, every team was solving the same problem independently. QuickBooks had its own task feature. Mailchimp had its own. TurboTax was building one. Each solution was slightly different, none talked to each other, and collectively they represented enormous duplicated engineering investment — and a user experience that felt incoherent.

The component system

I delivered a responsive component library designed to be resilient across the full range of Intuit's surface area — from narrow sidebar widgets (330px) to tablet views (768px) to full dashboard contexts (1024px).

Design Spec Responsive Sizes — 330 / 768 / 1024px

Task list is responsive based on the container it is housed in. Name truncates after two lines; description after three. Field ratio differs per size bracket.

I designed comprehensive resilience specs for Empty, Loading, and Error states across all components. The non-happy path is where design systems typically fail in production — building these specs explicitly ensured a consistently high quality bar across all 15+ Intuit SKUs.

Design Spec Component States — Empty / Loading / Error

Apart from the normal filled state, the task list defines explicit empty, loading, and error states — ensuring every team ships a consistent non-happy path without reinventing these patterns.

Design Spec Component Variations — default vs. status-aware

Task list variation depends on the variation of individual tasks within it. The left shows the default state — ghost CTAs. The right shows the status-aware state — TO DO, IN PROGRESS, and OVERDUE badges with filled CTAs that match urgency.

The design working model: solving the organizational problem

The hardest design problem wasn't the UI — it was governance. I designed the IPTM Design Working Model to answer how 15+ teams could build what their users need while maintaining platform coherence, defining three engagement scenarios:

Roadmap leadership

I drove the FY23 Development Roadmap — prioritizing the core services and widget library that would unlock the broadest adoption. For FY24, I defined the Discovery Roadmap focused on the next capability tier: Project Templates, Custom Fields, and Portfolio Health tracking for mid-market users.

Impact

Unified work management across 6 Intuit products and 15+ offering teams under a single design vision and operating model. Established the DACI operating model as the standard for how Intuit platform teams and offering teams collaborate on shared capabilities.