We hit this problem every day delivering AI-built software for clients
Running an AI-native services business since 2024 taught us where provenance and handover break. Gladiatar is the platform we built to solve this problem.
Jaskaran Sachdeva
PRODUCT AND GTM · LONDON
10+ years of experience across operations and product management
Ex-Uber
Early launch and Ops team at Uber UK Global Talent Visa Recipient for Exceptional Talent in Tech
Gaurav Chopra
ENGINEERING · BANGALORE
20+ years of experience in software development and engineering management
Ex-Amazon
Extensive experience in ML Ops; most recently technical Senior Engineering Manager at AWS Previously - Walmart Labs
1
client in production - 100% of PRs reviewed by Gladiatar
5
engineers shipping today
500+
Applicants to our AI Engineering challenge
gladitar· CONFIDENTIAL
03
The solution
Ship AI-built code faster, and prove it's safe.
What changes for your team.
Ship without the review bottleneck
Agents check every PR against the spec, so senior engineers review less and catch more.
How: review agent
3x more drift caught
before merge, live client work
Prove who built what, in minutes
Every change traced to its prompt, model and approver. Audit evidence on demand, not a month-long scramble.
How: provenance ledger
100% traceable
every change, end to end
Keep moving when people move
Decisions and intent travel with the code, so anyone can pick up the work.
How: documentation agent
Handover notes on every PR
written automatically
Know what AI costs, team by team
One predictable price, with spend shown by team and project.
How: spend view
$1 saves ~$15
of engineering time
Built for professional teams shipping production software, not for vibe coding.
gladitar· CONFIDENTIAL
04
How it works
Every stage has an agent and an owner. One record connects them.
Gladiatar orchestrates the flow, so work moves between people and agents without losing why it was done.
Delivery loop: production feedback updates requirements
Requirements
Ready
Production
Shipped
Design
NEXT
Code
PLANNED
Review
LIVE
QA
PLANNED
Document
LIVE
Deploy
PLANNED
Provenance ledger:every prompt, model, diff and approval
OUTPUT
Audit trail
Exported for auditors and regulators
GATE OPTIONS PER STAGE
Team member sign-off
Automated policy check
No gate
INSIGHTS
Cost and usage
GCU spend by team, agent and task
Quality signals
Where agents drift or fail
LiveNextPlannedConfigurable gate at every stageAny team member can own any gate
gladitar· CONFIDENTIAL
05
Demo
See it in action
Connect a repo
Code Review agent checks a PR against the requirements and flags intent drift
Documentation agent generates the API and User Feature documentation
Provenance view traces each change to its source
Spend view shows run cost by team
gladitar· CONFIDENTIAL
06
Traction
Already running on real client work.
Our engineers use Gladiatar on every release for our first client. Here's what changed.
$1 of Gladiatar saves ~$15 of engineering time
3x more drift issues caught before merge vs Standard Review
100% of changes traceable: prompt, model, approver
2 months in continuous use across 100+ PRs
Before
PRs reviewed by whoever prompted them
Context lost in chat windows
Docs written late, or never
With Gladiatar
Every PR checked against the spec
Handover notes written automatically
Full audit trail for every change
What's live
Review agent
Live
Documentation agent
Live
Provenance ledger
Live
Design agent
Next
Next: opening to 3 to 5 enterprise design partners in regulated sectors
gladitar· CONFIDENTIAL
07
Market
Every software team is becoming an AI-assisted team
$70bn+
TAM
Every professional developer worldwide
36.5M developers × $2000/yr
$10bn
SAM
AI-using engineers at enterprises in the UK, EU, US and India
~5M engineers, blended ~$2000/yr
$100m
SOM · ARR IN YEAR 5
~150 enterprises at ~$650k each
~500 engineers: ~$240k seats + ~$400k Gladiatar Compute Units
Revenue grows with every agent we ship, not just every seat we sell.
Sources: developer counts, SlashData. AI adoption (75%), SlashData Q1 2026. Spend validation: Cursor ~$4bn and Cognition ~$1bn run-rate, 2026 (Techcrunch). Seat mix and pricing are estimates.
gladitar· CONFIDENTIAL
08
Competition
Most AI agents optimise for output. Ours are built for accountability.
AI code review
Macroscope, CodeRabbit, Greptile
Augment
Cosmos, own compute
Gladiatar
Any agent, your record
Cognition
Devin, billed in ACUs
Factory AI
BYOM, own agents
Autonomous output ← → Human-owned and audited
Single stage ← → Full lifecycle
Humans and agents, one workspace: shared PRDs, reviews and approvals
Seamless handover: anyone picks up work with full context
Your workflow, your rules: custom stages, review standards and approval policies
Right model per task: chosen and managed for you.
gladitar· CONFIDENTIAL
09
Business model
Seats for people. GCUs for agents.
FROM $40 / SEAT / MONTH
Seats
For everyone who owns, steers or approves work. SSO, audit export and approval policies.
Grows with team size
PER AGENT RUN
GCUs
Gladiatar Compute Units. Priced by agent and task complexity.
Grows with every agent we ship
ANNUAL CONTRACT
Enterprise
Seats plus a committed GCU pool at a discount. Pay as you go above it. Paid pilots credited against it.
Private deployment available
New Agent = Incremental Revenue
From same customers
50% gross margin
blended target, models included
$50 to 100k → $1m+
land, then expand per customer
gladitar· CONFIDENTIAL
10
Go to market
Land with two agents. Expand with every new one.
Each customer grows three ways: more teams, more seats, more agents.
01 · Now to month 6
Design partners
3 to 5 enterprise teams in regulated sectors, plus our production client.
Goal: case studies and GCU usage benchmarks
02 · Month 6 to 18
First paying customers
5 to 10 paying customers through founder-led sales from London, in the UK and EU, with SOC 2 Type II in place.
Goal: seed-ready ARR
03 · After seed
After seed, scale
Enterprise sales hires in the UK, EU and US. Self-serve Team plan and cloud marketplaces as top of funnel. Every new agent expands existing accounts.
Goal: Series A, then $100m ARR by year 5
IDEAL CUSTOMEREnterprises with 200+ engineers in financial services, insurance and healthcare that use AI coding tools.
gladitar· CONFIDENTIAL
11
The ask
Raising $1m pre-seed
18 months to prove enterprise demand and ship the full agent suite, then raise seed.
55%
Product and engineering
India engineering team, infra and model costs for pilots
25%
Go to market
Founder-led sales, UK enterprise sales hire, content, events