Keops built Playbook six products in six weeks with Claude, saving $120k a year

Playbook designed and prototyped its next-generation platform with Claude, and validated it with investors before a line of production code existed. Working with Keops, that design became a live ecosystem: an AI assistant, a member matching engine, an operations back office and three client applications. A build of that scope conventionally takes a year but was delivered in 6 weeks. It unlocks $120,000 a year in annual technology costs and operational capacity, and turns a recurring software licence into an asset Playbook owns.

We went from renting our own community to owning it outright, and that took six weeks, not the year a build like this normally demands. The money we were paying out every year stays in the business, and what we built with it is worth something on our balance sheet.

Wafa AlObaidat, Founder, Playbook

A validated vision, and a platform that could not carry it

Playbook is a paid-membership community for professional women in the Gulf, founded in Bahrain by Wafa AlObaidat. Its growth had outpaced the systems its team ran on, and the community itself lived on a white-label platform, alongside separate subscription tools for ticketing and operations: together around $40,000 a year, every year, for software Playbook did not own and could not change.

Playbook had already done the hard product work. Using Claude, the team had shaped the product vision, designed the member journey end to end and built working prototypes, which it had taken to its investors. It knew what it wanted its members to experience, in detail, before it went looking for a build partner.

What the existing platform could not do was carry any of it. An AI assistant like Layla, or matching built on Playbook's own member data, were not things that could be added to someone else's product at any price. The gap was not the vision. It was the route from a validated prototype to a production ecosystem, and the time available to travel it.

We knew exactly what we wanted to build. We had designed the journey, prototyped it and shown it to our investors. What we needed was someone who could turn that into a real system fast, without watering it down.

Wafa AlObaidat, Founder, Playbook

Kickoff with Keops was 12 July 2026. Six weeks later the platform was built, in time to reach members at Playbook's annual Women Shaping Wealth summit in Riyadh, an event that draws around 3,000 attendees.

How the two teams split the work

The engagement worked because the division was explicit from the first week, and because both sides were already working with Claude.

That both teams used Claude is why the handoff cost almost nothing. Playbook's prototypes were not slideware to be interpreted, they were a working specification Keops could build against. Where a conventional engagement spends its first month converting intent into a document, this one started from something that already ran.

Pressure-testing the design against the systems

Keops began with Playbook's operations rather than its software. Working with Claude, the team mapped how Playbook actually ran, how members were vetted and onboarded, how programmes were scheduled, how introductions were made, and where data moved between systems, then tested the designed journey against what the underlying systems would have to do to support it. Where the design and the operational reality disagreed, Claude produced working interface variants within hours of a conversation, so the two teams resolved it on a screen rather than in a thread.

Claude with the context already in it

None of that worked from a blank page. Through MCP connectors, Keops gave Claude direct access to the systems where the engagement actually lived: the meeting note-taker that recorded every session with Playbook, the Slack channels where the delivery team worked, and the Keops knowledge base holding the process maps, specifications and decision records as they accumulated.

The effect compounds on a six-week timeline. A decision taken in a Tuesday call was available to whoever picked the work up on Wednesday, without a handover document. Questions that would normally cost a thread and half a day were answered from what had already been said.

What was built

Six products went from prototype to running software in six weeks, delivered by a Keops team of four.

Claude in the product, not only in the build

Claude Code compressed the build. The Claude API is what members actually use.

Layla runs on the Claude API with direct access to the platform's tools. She can read and update a member's profile, search the directory, request an introduction, enrol someone in a programme or answer a question about an event, because she is connected to those systems rather than describing them. A member asking who they should meet at the summit gets an answer drawn from their own profile and the live directory, in their own thread, on WhatsApp or in the app. It reads as a person who knows the community rather than as a chatbot.

Underneath, a second and much cheaper model does the quiet work. Claude Haiku reads the conversations members have with Layla and extracts the signals that matter: what someone is building, what they are looking for, what they can offer. Those signals feed the matching engine, so recommendations sharpen through ordinary use of the assistant rather than by asking members to fill in longer forms.

Nobody fills in a longer profile form. But they will talk to Layla. Haiku turns those conversations into the structured signal the matching engine needs, so the product gets better at introductions the more people simply use it.

Gabriel Morin, Founder and CEO, Keops

How a year became six weeks

Keops built with Claude Code, and the compression came from raw build speed. Features that would each have been a sprint were delivered in days: the AI assistant, the matching engine, the back office, the web application and both mobile apps, in parallel, by a team of four. The engineers stayed on the hard problems: the matching model, payments and entitlements, the mobile release path, and the data protection work.

Claude Code let us build at a rate that does not exist otherwise. Six products, six weeks, four people. Every one of those numbers is impossible on a conventional stack, and it is the same work, not a cut-down version of it.

Gabriel Morin, Founder and CEO, Keops

$120,000 a year, and an asset they own

The platform was complete six weeks after kickoff and went into testing with Playbook's own team from late August, on site at Playbook's offices, covering member vetting, programme scheduling, introductions, payments and both mobile applications. It is live now. Playbook is moving its membership across in waves rather than all at once: the first members are on the platform, building their profiles, using the AI assistant and sending feedback that goes into the product week by week.

The value shows up in two places, and they are different in kind.

That is $120,000 a year, and it changes the shape of the business as well as its cost base. Playbook owns the platform it runs on, holds the code and the member data, and can change any part of it without asking permission or waiting for someone else's roadmap. Spend that produced nothing durable has been converted into software on Playbook's own books: the recurring cost line becomes an asset, and the asset raises what the company is worth.

Our whole proposition is turning operating expenses into enterprise value. Playbook had a licence line funding someone else's product and a team spending its week on administration. Both of those go away, and what replaces them is an asset they own and people working on things that grow the company.

Gabriel Morin, Founder and CEO, Keops

What is next

Women Shaping Wealth takes place on 28 and 29 September in Riyadh, with Invest Saudi as Strategic Government Partner, bringing around 3,000 people together: members, investors and government entities. It is the best possible moment to put the platform in front of the whole community at once. Adoption that would normally build over months happens across two days, in a room where members can be walked through it in person.

It runs the other way too. The platform is one of the highlights of the summit in its own right, something members leave the event already inside rather than an announcement they read about weeks later, and a Claude-built product demonstrated in the heart of Riyadh to an audience of investors and government stakeholders.

Beyond the summit, the roadmap runs in the same direction: expansion in Saudi Arabia, deeper programme automation, and extending what Layla can do for members without a person in the loop.

The summit puts our entire community in one room, so it is the perfect place to hand them the platform, and it is one of the things I am most looking forward to showing them. We are not finished either. What we have now is something we can actually grow on.

Wafa AlObaidat, Founder, Playbook