We built Xlooop because customer trust dies quietly when AI ships unaudited work.
Three years ago every regulated firm worried about whether AI would replace them. Today they worry about something harder: proving to a regulator, an insurer, or a customer exactly what their AI did, who reviewed it, and what evidence backs it.
That gap — between AI velocity and audit-grade evidence — is the operating problem of the next decade. Xlooop is the layer that closes it.
Operator-built. Dogfooded in production governance work.
Xlooop is built on a multi-agent operating model that has been run in production governance work for over a year — concurrent specialist agents working in parallel, isolated by lease and branch so disjoint work never collides and overlapping edits are surfaced for human resolution, with hash-chained audit evidence.
full operator-evidence pack available under NDA · email hello@xlooop.com
How we work — every engagement, every release.
Audit-grade by default
Every AI output traces to the request that started it. Not a feature — a structural property.
Customer-respectful
We work for the firm. The firm’s customers see the same record their firm sees, no more.
Sterile by default
Zero connectors enabled at provision. Every data source explicitly approved. Teardown leaves no trace.
Model-agnostic
Use Claude, GPT, Gemini, local models. Switch any time. Your evidence trail survives the swap.
Vendor-neutral
Where equivalent connectors exist (Xero / MYOB / QBO; HubSpot / Salesforce / Pipedrive) we present them equitably.
Region-bound
Your data lives in your home edge. AU edge for AU; EU edge for EU; UK edge for UK; US edge for US.
The shape of what we won’t build.
- NOT CURSOR / COPILOT WORKSPACESwe are not a developer tool; we are the audit trail your customers, regulators, and team can read.
- NOT NOTION / ASANA / LINEARwe are not a PM tool; we are AI-aware and provable — to customers, regulators, and your own team.
- NOT DRATA / VANTAwe are not a SOC2 compliance scanner; we are AI-output-specific.
- NOT A UI BUILDERwe do not generate code; we govern AI work.
- NOT AI STRATEGY CONSULTINGwe are an operating layer, not advisory.
- NOT A TRAINING-DATA HARVESTERwe never train external models on your content.
Built by operators who lived the problem.
Australian-incorporated as ADEVI PTY LTD (trading as Xlooop). Global by default. Hiring across all five launch markets (AU · US · UK · DE · FR).
AI-governance · regulated SMB delivery · 15 years operating layer
TPB / OAIC / ICO / CNIL liaison · audit-grade evidence design
Cloudflare Access · per-tenant isolation · breach-response runbook
pilot graduation · workspace bootstrap · first-workflow discovery
individual names live on LinkedIn · contact hello@xlooop.com
Built by an operator who lived the problem.
Read the founder story →+
Marat Basyrov — Founder & CEO · Named inventor on US Patent 11,604,641
Marat has been operating in enterprise software delivery since December 2011, building and leading a Sydney + Melbourne-based consulting practice (Intelligent Profit Solutions; Adevi.io tech innovation, 2011–2022) across 130+ enterprise engagements (A$50K – A$2M ACV) before launching Xlooop. As Director – Product, Delivery & Solutions, he led enterprise consulting, pre-sales, solutions architecture, and offshore-team orchestration with Australia-based delivery leadership across health, financial services, public-sector, and regulated-industry deliveries.
The bi-directional code↔UI architecture and the AI-agentic delivery layer that underpin Xlooop's six-role panel were developed and validated across that 14-year operating window. The patent (filed April 2018 from Australian priority; PCT-filed April 2019; granted as US 11,604,641 on March 14, 2023) emerged from this body of work and covers the multi-user, role-based collaborative model and feedback-trained engine with chat-bot integration (Claim 1) — the foundation Xlooop's productisation is built on. Adevi IP Holding Company Pty Ltd holds the patent; ADEVI Pty Ltd (trading as Xlooop) is the operating company.
Marat dogfooded the operating substrate Xlooop is built on for 18+ months before any customer line of code shipped — a personal-life governance system spanning his own personal and corporate domains, refined across 100+ architecture iterations. Customers inherit a tested operating model, not a green-field experiment.
The personal-life-operations foundation underneath Xlooop:
- 38 active hard-rules — governance discipline applied consistently across personal and corporate domains
- 250+ specialist skills — reusable workflow primitives
- 100+ architecture iterations — audit-grade change records produced in real operating use
- Multi-agent collaboration registry — parallel-session conflict prevention; the underpinning of Xlooop’s hybrid configuration
- Lessons-learned + retrospective loops — the governance-learning feedback chain
Marat lived the AI-tooling-without-governance pain firsthand inside professional services and regulated SMBs. Xlooop is the product he needed and couldn’t buy.
Direct buyer empathy. Marat has personally carried Head of Delivery / VP Engineering accountability in mixed-tools, mixed-AI, multi-customer environments. The pain Xlooop solves is the pain he carried.
Contact: linkedin.com/in/maratbasyrov
Engineering co-founder + first engineering hire — Phase 1 hiring target. Full-stack engineer comfortable with Cloudflare Workers + audit-grade ledger patterns + multi-tenant isolation. Use of funds 35% allocated. AI agent team — see AI team.
Six convictions that shape every product decision.
Audit-grade is the moat
AI velocity without audit-grade evidence is borrowed trust. Audit-grade evidence — write-only, reviewer-signed, and provable to whoever asks (customers, regulators, insurers) — is the moat that compounds. Six months of provable records ≠ six weeks of churn. We design for it.
Hybrid beats either
Pure-human is slow. Pure-AI is unprovable. Hybrid is the most powerful configuration — humans gate authority; agents do scoped work; the audit trail is identical regardless of who or what produced the artefact.
Watch the tools your team uses
We don’t replace your stack. We watch it. Xero. MYOB. QuickBooks. HubSpot. Drive. Slack. Your AI tools. Your team. We produce the evidence trail that all of them, together, never quite manage to.
Authority-first, always
Your authority gate stays intact at every stage. Public-discovery first (L0). Sterile workspace next (L1). Scoped source access only when you explicitly grant it (L2–L3). Source-connected operations only when your governance is ready (L4–L5). The L0–L5 framework gives you a path; we follow your pace.
Vertical templates over vertical products
Generic workspace first. Vertical templates ship on top. We don’t sell you a different product because you’re a builder vs an accountant; we sell you the same product, customised by templates your peers built.
AU-first by deliberate sequencing — global by legal coverage
Australian-headquartered. Available worldwide. We comply with applicable law in Australia · the European Union · the United Kingdom · the United States. Initial GTM focus on Australian regulated SMBs (where the operator’s regulatory depth is strongest). The same generic workspace serves regulated SMBs globally. We are honest about regulatory depth — strongest in AU; adequate-at-SaaS-trust-level in EU + UK + US. This is the "global by legal · focused by GTM" pattern — the same pattern adopted by Stripe · Atlassian · Notion · Anthropic.
Our scope rule.
Generic workspace first. Vertical project spaces later.
This rule shapes every product decision. ASP + H&Y are early users of the generic workspace — not vertical-feature anchors. When the generic workspace is validated, vertical templates ship as drop-in starters; operator decides which verticals first.
Where we are. Where we’re going.
The audit-grade evidence we promise — to customers, regulators, and your own team — begins with the audit-grade evidence we keep about ourselves. "What’s authored ≠ what’s shipped."
What we’ve built
A working generic-workspace shell + canvas + 6-role panel + audit-trail primitives + 250+ specialist skills + multi-agent collaboration registry + 100+ architecture iterations dogfooded in real operating use.
What we’ve sold
Zero paying customers as of May 2026. Two early-user pilots engaged (ASP + H&Y; public-discovery stage; draft consent posture pending principal review).
What we’re building next
The marketing website you’re reading; production hosting on Cloudflare Pages; the pilot conversion engine; first paid pilot conversions.
Pick the one that matches your readiness.
5 minutes · 30 minutes · or a short message on linkedin.com/in/maratbasyrov
Three doors. Same destination: a firm whose AI work is provable.
Pick the one that fits where you are today — none commit you to anything.
Or simply email hello@xlooop.com — one human reads every message.