SonicJobs Wants to Kill the Redirect - the 12 Seconds Where Job Hunts Go to Die
Ninety-five percent of job seekers quit before they finish applying. SonicJobs built an AI agent to do the quitting-worthy part for them - and turned the recruitment industry's worst funnel into its pitch.
Every online job hunt has the same silent cliff. You find a role you like on a job board, you tap "Apply," and then - instead of applying - you are thrown to a different website you have never seen, asked to make an account, and told to re-type the resume you already uploaded. Somewhere in those 12 seconds, most people close the tab. SonicJobs was built around that exact moment, and around a number the recruitment industry has quietly tolerated for years: roughly 95% of candidates who start an application never finish it.
The London- and San Francisco-based company, founded in 2017 by Mikhil Raja, Francesco Masia and Francesca Boccolini, does not try to out-list Indeed or out-network LinkedIn. It goes after the plumbing. SonicJobs builds AI agents that complete a job application in place - reading the form, understanding each field, answering the screening questions, and submitting - without redirecting the candidate anywhere. The company frames the whole effort in a phrase that is deliberately unglamorous: it is "building Shopify for Jobs," infrastructure that other platforms switch on, not another destination fighting for your attention.
01 / THE PROBLEMA funnel that leaks 95 of every 100 people
To understand why SonicJobs exists, you have to look at the shape of the online recruitment market rather than any one app. In the United States alone there are more than 8 million job openings a year, roughly a million employers, and - crucially - over 300 different applicant tracking systems (ATS) sitting behind those jobs: Workday, Greenhouse, iCIMS, Taleo and hundreds more. Each one has its own login, its own form, its own idea of how many times you should type your phone number.
A candidate browsing a job board rarely sees that machinery. They see a listing and an "Apply" button. What happens next is the redirect - the hand-off from the board to the employer's ATS - and it is where the funnel collapses. SonicJobs cites a bounce rate of around 70% at that step. Compound that with account creation and repeated data entry, and the industry-standard completion rate settles at about 5%. In plain terms: for every 100 people who want a job enough to click, roughly 95 walk away before finishing.
That fragmentation is not a bug the incumbents forgot to fix; it is the water everyone swims in. SonicJobs' bet is that the leak is so large and so universal that closing it is worth more than any new feature bolted on top.
02 / THE FIXTeaching an agent to fill the form for you
SonicJobs' answer is "Hosted Apply": the application flow is hosted in place, so a candidate completes and submits without ever leaving the page they started on. Behind that simple experience is what the company has been quietly building since 2019 - AI web agents that combine robotic process automation (RPA), computer vision and large language model reasoning to interpret an application form the way a person would, then act on it.
The technology matters here because job applications are unusually hostile to automation. Forms vary wildly, fields are labelled inconsistently, and screening questions can be conditional. Early generations leaned on RPA and machine learning to recognise fields. The latest - the company's eighth generation, launched in 2025 - runs on OpenAI's Computer Use models, letting the agent autonomously navigate forms, verify a candidate and handle screening across multiple job sites. Notably, this requires no engineering work or integration cost from the employer or the ATS.
Roughly a 5x lift in completed applications. The same top-of-funnel traffic, far more people reaching the finish line.
03 / GOOD FRICTION, BAD FRICTIONThe idea underneath the product
The most transferable thing about SonicJobs is not a model; it is a distinction the team keeps returning to. There is "bad friction" - re-typing your address for the fortieth time, creating an account you will never revisit, uploading a resume the system then ignores. And there is "good friction" - the genuine screening questions that help an employer decide whether you fit. The product deletes the first and keeps the second.
That framing is why SonicJobs is careful to say it delivers "high-quality applications that lead to hires," not just more applications. An auto-apply tool that floods employers with low-intent submissions solves the candidate's problem by creating the employer's. SonicJobs is trying to sit on both sides of that trade - which is also why its customers are the platforms and employers, not only the job seeker.
04 / HOW IT'S DIFFERENTInfrastructure, not another auto-apply bot
There is a growing shelf of consumer "auto-apply" tools - Simplify, LazyApply, Teal and others - that install as browser extensions and spray a candidate's resume across hundreds of listings. SonicJobs is a different animal. Instead of acting on behalf of one job seeker against the platforms, it sells completed-application infrastructure to the platforms and employers themselves, then hosts the flow so candidates finish in place.
| Job boards (Indeed, LinkedIn) | Auto-apply tools | SonicJobs | |
|---|---|---|---|
| Where you apply | Redirect to ATS | Extension on top of ATS | In place, hosted |
| Who's the customer | Employers (job ads) | The job seeker | Platforms & employers |
| Optimises for | Clicks / listings | Volume of applies | Completed, quality applies |
| Integration cost | - | None (browser) | None (no dev work) |
The commercial logic follows the design. Because there is no API project to run, partners can go live in days rather than months. In one partner trial, hosting the flow cut the effective cost-per-application by up to 67%. That is the sort of number that gets an infrastructure company into rooms full of employers.
05 / THE MARKET & THE MONEYWhere SonicJobs sits
US talent acquisition is a market of roughly $15 billion a year, and SonicJobs is positioning itself as connective tissue across it rather than a single node. Its business model is B2B2C: the job search app is free for candidates, while employers, ATS vendors and job platforms pay for high-quality completed applications and license the Hosted Apply layer. Partners so far include the AI resume startup EarnBetter and the marketplace OfferUp, with more than 100 business customers reported by 2025.
The company raised a $5.5 million (£4.2 million) Series A in December 2021, led by Triple Point Ventures, TMT Investments and Ufi Ventures, with a long roster of angels from Sky Bet, PropertyGuru and JetSynthesys among others. The round funded deeper machine learning and the US launch. Its recognition has since come less from headlines and more from the industry itself: membership of NVIDIA's Inception program, a feature on the NVIDIA AI Podcast, and, in September 2025, first place at the HR Tech Pitchfest in Las Vegas, beating out more than 100 applicants.
06 / THE FOUNDERFrom Deutsche Bank to the apply button
Mikhil Raja is an unlikely candidate to obsess over form fields. Born in Mumbai and a competitive junior tennis player before an economics degree at Edinburgh, he spent close to a decade in investment banking - a stint at Rothschild, then years at Deutsche Bank - before committing to SonicJobs full time as CEO. His co-founders round out the picture: Francesco Masia, the CTO with a computer science background from Turin, and Francesca Boccolini, the CMO who studied technology entrepreneurship at UCL.
Raja tends to talk about the work in stakes larger than software. It is a useful reminder that the thing being optimised - the completion of a job application - sits at the front of decisions about income, careers and families.
07 / WHAT YOU CAN TAKE FROM ITFind your redirect, then delete it
Strip away the recruitment specifics and SonicJobs is a case study in a repeatable move: find the single step in a funnel that everyone has accepted as normal, prove how much it leaks, and build the thing that removes it. The company did not invent a new job board; it found the 12 seconds where hunts die and pointed an AI agent at them. It was also early on purpose - building application agents from 2019, so that when foundation models finally caught up, it had a multi-year head start and real-world data to fine-tune on.
The honest caveats are worth stating too. An agent-completed application only helps if the quality holds up on the employer's side, which is why "good friction" stays in. And an infrastructure business lives or dies on distribution partnerships - platforms and ATS relationships - more than on any single clever model. For SonicJobs, the model is the easy part. The redirect is the enemy, and the whole company is the argument that it never should have existed.