Why We Built Rankr: The Story Behind Our AI CV Screening Tool
Every startup has an origin story. Ours started with watching recruitment agency owners spending their evenings reading CVs when they should have been at home with their families. Here's why we built Rankr, what we learned, and where we're going next.
The Problem We Saw
Before building Rankr, I spent time talking to recruitment agency owners around the UK. The same pattern kept emerging: successful agencies were drowning in their own success.
They'd win new clients, post job ads, and get flooded with applications. Great problem to have, right? Except someone needs to read all those CVs. And when you're a small agency with 3-10 recruiters, that someone is usually the owner, working until 10pm to get through the pile.
One agency owner told me: "I love recruitment. I hate spending 60% of my time reading CVs instead of actually talking to people." That stuck with me.
Why Existing Solutions Didn't Work
The obvious question: aren't there already CV screening tools? Yes, there are. But they had three main problems for small UK agencies:
Too Expensive
Enterprise recruitment software with AI features typically costs £5,000-15,000 per year minimum. For a small agency doing 50-100 placements annually, that's a significant chunk of profit. The ROI wasn't obvious enough to justify the cost.
Too Complex
Many existing tools require extensive setup: training the AI on your specific requirements, integrating with your ATS (if you even have one), weeks of configuration. Small agencies don't have time for that. They need something that works immediately.
Not UK-Focused
Most recruitment software is built for the US market. Data stored in US servers (GDPR nightmare), pricing in dollars, features optimised for American hiring practices. UK agencies felt like second-class customers.
Our Approach: Simple, Affordable, UK-First
So we decided to build what the market was missing: CV screening software specifically for small-to-medium UK recruitment agencies.
Simple
No complex setup. Upload CVs, describe the role in plain English, get ranked results in 5 minutes. If you can use email, you can use Rankr.
Affordable
We priced it at £39-119/month for most agencies. That's less than your monthly coffee budget, not a major capital investment requiring board approval.
UK-Focused
Data stored in UK servers, GDPR compliant by design, pricing in pounds, built by a team that understands UK recruitment. We even spell "realise" correctly.
Technical Challenges We Solved
Building CV screening software is harder than it looks. Here are some of the problems we had to solve:
Reading Messy CVs
Candidates submit CVs in every format imaginable. PDF, DOCX, weird margins, tables, columns, funky fonts. Our AI needed to accurately extract information from all of them. This took months of testing and refinement.
Understanding Context
Early versions of our AI were too literal. If a job required "project management" and a CV said "led cross-functional initiatives", it wouldn't make the connection. Teaching AI to understand semantic similarity rather than exact matches was crucial.
Explainable Scores
Just giving someone a score isn't useful. Recruiters need to know why someone scored 87% vs 62%. Building explanations that were actually helpful (not just "AI magic") was surprisingly hard.
Speed
Our target was screening 100 CVs in under 5 minutes. Early versions took 20+ minutes because we were making too many API calls to the AI model. Optimising this without sacrificing accuracy was a technical challenge, but we got there.
What We Learned from 20+ Agencies
Before launching, we partnered with over 20 UK recruitment agencies to test Rankr. Here's what we learned:
Agencies Don't Want Another Platform
They already have an ATS, email, LinkedIn, job boards. They don't want to log into yet another system daily. Our solution: make Rankr work as a standalone tool that doesn't require constant use. Upload CVs when needed, get results, done.
Trust Takes Time
Recruiters were initially sceptical of AI. "Will it miss good candidates?" Fair concern. We addressed this by being transparent: show match scores with explanations, let recruiters see all candidates (not just top picks), make it easy to override the AI's ranking if you disagree.
Volume Varies Wildly
Some agencies screen 20 CVs per week, others screen 500. We needed pricing tiers that work for both without penalising small users or leaving money on the table with heavy users.
What We're Building Next
Rankr is just the beginning. Based on agency feedback, here's what we're working on:
ATS Integration
Many agencies want Rankr to pull CVs directly from their existing ATS and push ranked results back. We're starting with the most popular UK systems: Bullhorn, Vincere, and JobAdder.
Auto-Generated Screening Questions
If the AI identifies potential concerns about a candidate (e.g., skill gaps, unclear experience), it could automatically suggest screening questions to ask. This would save recruiters time figuring out what to probe on in interviews.
Team Collaboration Features
Right now, Rankr is single-user focused. Agencies want the ability to share rankings with colleagues, add notes, and collaborate on shortlisting decisions. Coming soon.
Being Honest About the Journey
I'll be honest: building Rankr has been harder than I expected. Technical challenges, customer acquisition, pricing strategy, competition from bigger players. There were moments of doubt.
But every time an agency tells us Rankr saved them 15 hours last week, or helped them find a candidate they would have missed, it validates why we're doing this.
We're not trying to be the biggest recruitment software company. We're trying to build the best CV screening tool for UK agencies who need something simple and affordable that just works.