You can't test AI-speed code with human-speed tools.
Our Solution
We test like humans see. Not like robots parse.
AI agents that see your app like users do.
DOM-Based Tools
Parse HTML structure
Target selectors (IDs, XPaths)
Break when code changes
X "Click #btn-42"
Pie
See the actual screen
Target visual elements
Work if UI looks the same
O "Click the blue Add to Cart button"
Vision
Tests what users see
0
Code Required
Auto
UI changes? Pie adapts.
All
One brain. All platforms.
Product
80% coverage in 30 minutes. Every time.
Step 1
2 min
Connect Point Pie at your app. No SDK.
Step 2
10 min
Discover AI agents explore and map flows.
Step 3
15 min
Generate Plain English test suites.
Step 4
Continuous
Self-Heal UI changes? Pie adapts.
What a real test run looks like:
O 142 tests passed
? 3 tests self-healed (UI changed, logic didn't)
X 1 real bug found: "Checkout fails when cart >10 items"
- Screenshot + video attached
- Suggested fix: pagination overflow
False positives: 0
Proof
Enterprise-validated. Fortune 500 in active pilot.
$200K
ARR
$90K
Largest ACV
0%
False Positives (90 days)
F500
In Pilot
Tilt (Fintech) - $90K ACV
Before: 4-6 hrs manual regression/release
After: Zero manual intervention to App Store
Expanding: 1 app → 4 apps
Won: Beat competitor at $60-70K on tech
Fi (Consumer HW) - $36K ACV
Before: 2-3 days QA, 12+ engineers
After: Hours, not days. 3 people.
Expanding: Mobile → e-commerce web
CEO approved: Case study signed off
DICK'S Sporting Goods - Fortune 500 Pilot
Market Cap
$16.8B
Status
Through Security
Test Cases
300+ importing
Impact
Kills concentration
Timing
Three forces converging. One window.
1. AI coding hit the enterprise
1.8M+ Copilot subscribers
Cursor: Fastest-growing IDE ever
40%+ of commits AI-assisted
That wall is us.
2. Legacy QA is broken
2000-2020
Selenium
Worked
2020-2024
Low-code
Struggling
2025+
???
Broken
Can't be patched. Needs new architecture.
3. Money moved to applications
Foundation models → Application layer → High-value workflows.
Pie is the application layer for the most broken workflow in software.
The window is now.
Market
$50B market. Actively breaking.
Software Testing & QA
$51B→$107B
11.3% CAGR
2025 → 2032 (Markets and Markets)
Test Automation
$35B→$77B
16.8% CAGR
2025 → 2030 (Coherent Market Insights)
$15B
Enterprise QA Tools
Primary Target
$5B
AI DevTools
Emerging
$20B
QA Services
Replacement
Competition
The only vision-based agentic testing platform.
Capability
Pie
BrowserStack
Mabl
Rainforest
Vision-Based
O
X
X
X
AI-Refactor Resilient
O
N/A
X
Partial
Cross-platform
O
Infra
Web
Limited
Self-healing
O
X
O
O
The flywheel they can't replicate
1
Data Flywheel: Every test run trains our models
2
Cross-Platform Memory: Stateful understanding. Not just vision.
3
Production Gate Lock-In: Ripping us out is politically expensive
By the time they rebuild, we own the data and the relationships.
Business Model
Usage-based. Indexed to AI velocity.
Starter
$500
/month
Up to 1,000 test runs
Growth
$2,000
/month
Up to 10,000 test runs
Enterprise
Custom
Avg $60-100K ACV
Unlimited + SLA + Support
The Flywheel
Customer adopts AI coding →
Ships more features →
Runs more tests →Pie revenue grows
We're the tax on AI development velocity
Expansion Evidence
Tilt: 1 app → 4 apps (4x potential)
Fi: Mobile → Web (cross-platform)
Both expanding within 6 months
Team
Built by engineers who lived this problem.
Dhaval Shreyas
CEO
Square → Instacart → Meta. CTO/Co-Founder @ Cue. M.S. CS, CU Boulder. Built mobile at Square. Saw QA become the bottleneck. Had to fix it.
Jinoo Jain
CPO
OpenSpace → Nightingale (drones). NASA Ames; UC Berkeley Robotics. Shipped AI that sees the world. Pie applies vision to testing.
Adithya Aggarwal
CTO
M.S. CS (ML), ASU. ACM published. Built face recognition & visual ML. Career in visual recognition. Testing should work like humans see. Built the engine.
~18 people. Heavy R&D = building the moat.
Engineering (14)
Product (2)
GTM (2)
Proprietary vision engine. Built from scratch. Not a GPT-4V wrapper.
Enterprise DNA. Closed $90K against cheaper incumbent. We know how to sell.
The Ask
$3.5M to own the verification layer.
Round Details
Raising
$3.5M - $4M
Valuation
$18-22M Pre
Instrument
Priced / SAFE
Timeline
Q1 2026
Why this valuation: AI DevTools median $18-25M pre. Proprietary vision engine, F500 in pilot, both customers expanding.
Use of Funds
Sales & Marketing50%
R&D35%
G&A15%
18-month milestones → Series A ready
$200K →
$1.5-2M
ARR
5 →
15-20
Customers
45% →
<20%
Concentration
18 →
30-35
Team
18 months runway . Seeking lead with Agentic AI / DevTools thesis
The immune system for the AI-native software stack.
Every company shipping AI-generated code needs verification that can keep up.
We built it. We proved it. Now we're scaling it.
The question isn't whether this market exists. It's who owns it.
