AB Income Architect - Retirement Income Planning Platform

We designed and built Alpha Beta Income Architect for Alpha Beta Partners, turning an academically grounded retirement-income model into a multi-tenant SaaS platform for UK financial advisers. Our work covered product design, the adviser application, organisation administration, reporting, integrations and the infrastructure connecting the retirement-income modelling service to the web platform.
The academic foundation came from Professors Andrew Clare and Stephen Thomas. Our role was to make that methodology usable in everyday advisory work: an adviser can configure a client's plan, test scenarios, explain projected income and remaining capital, and return to that plan as the client's circumstances change.
We are responsible for
Challenge
Alpha Beta needed to make a specialist retirement-income model easy for advisers to use and its results easy for clients to understand.
The methodology examines sequence-of-returns risk: the order of investment returns matters when a client is withdrawing from a portfolio. Poor returns early in retirement can have different consequences from the same returns arriving later. The product had to make those consequences understandable through income projections, capital balances and explanations of uncertainty.
Early design reviews showed that advisers needed to test scenarios during a client conversation: change an allocation, examine the effect, and explain the trade-off without losing their place. Confidence bands, which show annual withdrawal estimates at different levels of caution based on historical market outcomes, also needed a clear explanation so clients could understand the trade-off between drawing income and preserving capital
For UK firms working within FCA suitability and Consumer Duty expectations, clear interpretation and justified assumptions matter as much as the calculation itself. The FCA's cashflow-modeling guidance reinforces that context. Alpha Beta also needed to operate the product across multiple firms, with separated client records, controlled configuration and a record of changes and reviews.
Solution
Redberry designed and built a Laravel-based SaaS platform around Alpha Beta's retirement-income methodology. The central product decision was to bring plan inputs and Quick Insights into one workspace, so advisers could adjust a scenario and discuss its consequences with a client without moving between separate entry and analysis screens.
We made interpretation part of the design: named income profiles give the confidence bands a practical context, inflation explanations distinguish purchasing power from cash amounts, and Client View and PDF reports provide a clearer format for discussion. Behind the interface, Laravel manages plans and firm configuration while a separate Python/FastAPI service performs the specialist calculations.
Each advisory firm has its own tenant database, permission-controlled settings and branding. Multiple pots, estimated tax, blended income and intelliflo integration extend that foundation into the adviser's working process. The result connects the research model to a usable, continuing planning workflow, with projections presented as estimates based on data and assumptions.
We re-implemented Alpha Beta's academic model in Python for production use, preserving its analysis of historical return sequences while building the data-import and configuration workflows around it.
Advisers can use Alpha Beta's supplied portfolios or their firm's other preferred models, with historical performance data imported into the platform. Firm-level model selection controls which options advisers can use. Laravel assembles the plan and portfolio inputs for the modeling service, then brings income and remaining-capital projections back into the interface.
Selectable confidence bands and real and nominal views let advisers explore the results and distinguish purchasing power from cash amounts. The historical inputs give that conversation an inspectable basis; the projections remain estimates rather than guaranteed outcomes.
We designed the planning workflow so advisers could change inputs and examine their consequences in the same workspace.
The React and TypeScript interface keeps the planning controls on the left and Quick Insights on the right. Changes to relevant parameters pass through Laravel to the Python modeling service, and the returned results update the income summary and capital projections. Missing-data guidance explains what is needed before a projection can be produced.
That combination supports a concrete advisory task: adjust an allocation or income source, see how the plan changes, and explain the trade-off while the inputs remain available. The interaction follows the client conversation rather than the structure of the research model.
We designed the results presentation to help advisers explain the model, as well as operate it. Named profiles frame confidence bands around objectives such as preserving capital, balancing income and capital, or drawing more income. Annotations explain uncertainty, inflation and purchasing power beside the figures.
The adviser-facing Client View brings the retirement snapshot, income composition and projected balances into a simpler discussion format. We also built organization-branded PDF exports that store a snapshot of the selected plan, calculation data and confidence view.
That snapshot preserves what was captured for a particular report as the working plan evolves, giving the adviser a stable reference for sharing results and returning to them at a later review.
We built the plan data model and calculation flow to represent separate retirement pots and their tax treatment. A pension and an ISA can have their own balances and allocations, with taxable and non-taxable income handled separately.
The tax layer applies configured allowances and bands to estimate liability and net income. Gross and net views connect the projection to spendable retirement income: the same withdrawal amount can leave different amounts available after tax, depending on how it's treated.
We built the multi-tenant foundation with a separate database for each advisory firm's tenant records, using Stancl Tenancy for provisioning and tenant-aware services. Central administration maintains shared product foundations and defaults.
Permission-controlled workflows let authorized firm users manage users, available models, fee tiers, income sources and standard inflation, wage-growth and RPI assumptions. This makes ownership of the standard assumptions explicit while preserving the adviser’s client-specific planning choices.
Firms can customize the application’s look and feel with their own branding, including logos, interface themes and chart colors, and apply the same visual identity to PDF reports. One product can therefore support distinct firms' records, administration and presentation.
We built the records around the projection to support the ongoing adviser-client relationship. The dashboard highlights upcoming annual reviews; client and plan notes retain discussion context; plan states distinguish working alternatives from active and archived plans.
Activity records capture who changed a value, when, and the relevant before-and-after values. Together with report snapshots, this gives advisers a practical reference for the next review: what was presented, what was recorded around it, and what has changed since.
We implemented the linked-app launch and data-read workflow that connects Income Architect to a firm's client context in intelliflo.
The launch resolves the organization, adviser and client, with request verification and tenant/user matching at that boundary. Connected APIs supply relevant client and contact information to the planning journey. This lets advisers bring the specialist income model into their existing workflow, with the client context already available when they begin planning.

We are a 200+ people agency and provide product design, software development, and creative growth marketing services to companies ranging from fresh startups to established enterprises. Our work has earned us 100+ international awards, partnerships with Laravel, Vue, Meta, and Google, and the title of Georgia’s agency of the year in 2019 and 2021.



Mailbox for Laravel, our open-source package for capturing and previewing a Laravel app’s email, reached version 3.0 on October 6. It is a security and stability release. The dashboard is now locked to local development by default, email previews are sanitized more safely, and the dashboard loads faster. Teams that open Mailbox on staging need to add one gate before upgrading.
.webp%3Fv%3D1793872777&w=3840&q=75)
The second edition of Laravel Meetup Georgia is on Thursday, October 8, at 18:00, on the Stamba rooftop in Tbilisi.

Redberry is twelve. What began in 2014 with four founders in Tbilisi now rests on two decisions taken along the way: specialise in Laravel, and take that expertise international. That adds up to 300+ products shipped across the EU, UK, US and Middle East, around 200 people, and a place among the 13 Laravel Premier Partners worldwide. Here is how we got here, and who got us here.