An AI mission-control deck
Runs and coordinates 50+ live AI agents and 17 AI project leads across my companies, each lead owning an ongoing business or engineering goal.
Built since June 2026A private AI building circle for founders
A private, hands-on circle with Dien Limano for founders of established companies. See how I build, bring your own challenges, and work alongside me on systems that matter to your business.
Who can apply
For founders running companies with roughly US$170K+ in monthly revenue. Ten founders maximum, including Private members. Personal participation required.
Applications are not open yet.
~20
business systems built or adapted, running in production
50+
live AI agents coordinated from one mission-control deck
67,000+
customers on systems I built
84,000+
customer conversations on systems I built
How it works
Orders, invoices, chats, stock counts and ad numbers pile up in different places and different shapes.
I built or adapted about 20 systems that tidy sales, finance, stock, marketing and customer data.
The systems are joined up, with a single sign-in, so the data can be reached from one place.
An AI desk that runs 50+ agents works across those systems. The top of the page shows a broadcast and a page test; four more:
Checks the books every working day, traces anything that doesn’t add up, and asks the head of finance before a single entry changes.
An AI team writes, animates and schedules short videos across six platforms; I only pick the ones I like.
Checks chat bug reports against the app’s records, fixes the real ones, has a second AI audit each fix, and ships after my go. One sweep: 14 reports, 3 real bugs.
Sends each chat enquiry to the teammate most likely to close it. Weekly weight updates need my OK, and it falls back to round-robin if results slip. About 5,200 routed since 6 Oct.
Better decisions, acted on sooner. The goal is more profit for the company.
An illustration of the approach. The broadcast, bug-sweep and enquiry figures are real; the pictures show how the work runs, not company data.
Who is building
I build the AI systems that run my own companies. My work spans customer conversations, CRM, accounting, social publishing, creative tools and my own AI mission-control deck.

Some systems are in production; others are experiments in progress. I show the difference, including what failed and what I am still learning.
17
AI project leads, each owning an ongoing goal
11,000+
commits in the past 12 months across 25+ apps and systems
~5,000
AI ad creatives generated, analysed and published for paid campaigns
< 1 week
native iPhone and Android apps, from first commit to a working build
Described by what each system does. Some are built from scratch, some adapted from open-source platforms.
Runs and coordinates 50+ live AI agents and 17 AI project leads across my companies, each lead owning an ongoing business or engineering goal.
Built since June 2026Self-hosted shared inbox with sales-lead routing, attribution and sales reports. Adapted from an open-source platform.
84,000+ conversationsCustomers, commerce, campaigns, ads, content and support in one place, with a single sign-in for the other internal apps.
67,000+ customersGenerates, analyses and publishes ad creatives for paid campaigns.
Nearly 5,000 AI creativesDocuments, financial reports, stock, tax and payroll, built to replace the previous accounting software.
Used by the finance teamMakes routine automated judgements for the agents. It beat the commercial model it replaces on our internal benchmark.
88.6% vs 74.6%These systems hold real customer and company data, so I attack them before anyone else can, on safe copies.
Testing
In place
Each week, I show something I am building or working through. You ask questions, bring a relevant problem, and see how I would approach it. When the session involves a build, you work on your own laptop at the same time.
Three hours, live, every week
Between sessions you try things and bring back what worked, what failed and what you want to improve. You do not need to arrive as a programmer. You do need to participate personally. A default shape, not a rigid syllabus.
Neither includes done-for-you development, unlimited debugging, emergency support, maintenance or security certification. You and your team keep ownership of implementation.
For
Not for
No programming background is required. You must be willing to build personally, ask questions, test and troubleshoot.
The seat belongs to the founder. Your team can support implementation, but cannot replace you in the Circle.
No. The programs provide live examples, advice and guided work. Done-for-you development and maintenance are outside the scope.
Circle membership does not include personal direct-message coaching. Private AI Partnership includes priority messaging with an agreed response target.
Your own laptop and an AI coding subscription, ideally in the US$100–200 a month range. It’s paid by you, separately from the membership, and it’s what lets you build alongside me.
No. Each term is three months, and that can be enough to learn a lot and carry on by yourself. If you see how much you are growing and want to keep going, you can renew for another term.
Show me before our second Circle meeting. In a short interview you walk me through what you had built before joining. If it goes beyond mine, I’ll refund your full first payment plus 10%.
Admission will be by application, reviewed personally. Ten founders maximum.