Ten Relevant People. One Focused Daily Habit.
A private, human-in-the-loop prospecting workspace that discovers and ranks useful professional connections without automating LinkedIn activity.
View live projectRanked around each user's goals
Explainable evidence and confidence
Next.js and React Native clients
Workspace and email preview
The challenge
Professional networking tools often optimise for volume. Daily Ten needed to do the opposite: create a small, deliberate habit that helps a user find useful people without sending automated messages, connection requests or other actions on their behalf.
Relevance had to be understandable. A black-box recommendation list would not give users enough confidence to act, so every candidate needed a score, evidence, confidence signals and a concise explanation connected to the user's role, goals, expertise and target audience.
The daily promise also created an operational challenge. Public-web discovery can be variable, while scheduled lists and emails must remain reliable across time zones, provider delays and multiple accounts without producing duplicates or exposing one user's data to another.
The approach
Personalised onboarding
A five-stage onboarding flow captures the user's professional context, networking goals, ideal audience, exclusions and preferred delivery schedule, creating a usable targeting model from the first session.
Explainable daily ranking
A deterministic 100-point model applies hard exclusions, confidence and evidence fields before selecting ten candidates plus ranked reserves. Feedback and confirmed outcomes improve later recommendations without hiding the reasoning.
Resilient discovery and delivery
Apify public-web discovery, an optional Apollo fallback and a rolling qualified-candidate pool keep generation away from the email-critical path. PostgreSQL leases, durable job records and idempotent queues protect scheduled delivery.
One product across channels
The Next.js workspace, daily email preview and Expo React Native client share the same accounts, queues and API. Users review candidates, open LinkedIn themselves and explicitly record what happened.
The results
Daily Ten became a production private-beta SaaS product with secure accounts, personalised targeting, daily ranked queues, searchable history, outcome tracking, insights, scheduled email delivery and an administration layer for monitoring recommendation quality.
The product is deliberately human-led. It reduces the work required to identify promising connections while leaving every LinkedIn action, note and outcome under the user's control. No invented adoption or revenue claims have been added to this case study.
What did not go perfectly
The first scheduling design tried to discover candidates during the morning generation run. Provider timeouts could leave the previous day's list in place, so discovery was separated into a rolling background pool and routine delivery became a fast, deterministic selection step with a clearly labelled partial-list fallback.
Project gallery
Screenshots and project visuals from the build, optimisation, or platform work.
Services delivered
Related case studies
More examples across ecommerce, platform work, and custom development.
Mission Control AI App Scanner
A secure software-assurance platform that grades ZIP uploads and GitHub repositories, explains every finding, and identifies AI app-builder signals with confidence and evidence.
Read case study →
Sustainability Assessment Platform Custom App
Custom web application development for a sustainability assessment workflow that needed structured data, collaboration, auto-save, and report generation.
Read case study →
4x4AT Custom Ecommerce POS
A PHP-based staff sales application built for Magento and adaptable to other ecommerce platforms, supporting call-centre, trade and B2C order workflows.
Read case study →Got a project like this?
Book a free 30-minute Discovery call. We will talk through what is possible for your business and whether I am the right person to deliver it.