How this plays out in real engagements.
What we found, what we proposed or built, and what we can honestly say about the result so far.
We only publish work here once we can describe it accurately. Some of it is finished and measured. Some of it is a proposal awaiting a decision, or a build still in progress. Each entry below says which, and what the actual status is today.
Designing a measured path from a broken ATS to an AI-native recruitment platform
UK recruitment and staffing client. Name withheld by agreement.
The Opportunity
A UK recruitment and staffing business had outgrown the in-house platform it built to run its desks. Candidate data was duplicated and unreliable, search barely worked, and screening and candidate communication were manual, which left recruiters spending their time on admin instead of placements. The system of record was no longer trusted enough to run the business on, and a rebuild had to solve the trust problem, not just add features.
Our Approach
Printmind read the business before proposing a solution. That work produced AI-RecruitOS: a modular, multi-tenant platform pairing a reliable ATS core (jobs, candidates, pipeline, submissions) with an AI layer for CV parsing, candidate screening and ranking, natural-language rediscovery of dormant candidates, and automated engagement.
Delivery is structured as a phased, gated roadmap, with each phase tied to defined success KPIs, including time-to-shortlist, screening time per role, CV-parsing accuracy, recruiter adoption and cost-per-tenant, so build investment is only released against measured results. The engagement also produced the commercial foundation needed to productise the work: per-tenant run-cost modelling, a pricing approach, and an IP and licensing structure giving the client a path to a bootstrapped SaaS offering.
The Impact
The product definition, roadmap and KPI framework are complete and are currently under client evaluation ahead of a Phase 1 build. Structuring success criteria before writing any code means that once Phase 1 goes live, impact will be measured against a baseline set in advance, not asserted after the fact.
Bringing DRS-grade decisions to the other 99 percent of cricket
Howzzat.ai, a Printmind venture.
The Opportunity
Cricket is played by an estimated 300 million people worldwide, the vast majority at grassroots and amateur level: local leagues, turf grounds, academies and weekend clubs. At this level, every match relies entirely on human umpires making split-second decisions on wides, LBWs, run-outs and edges, with no technology assistance at all. Wrong calls are frequent, and in competitive matches they damage team relationships and tournament credibility.
The professional solution, DRS powered by Hawk-Eye, costs hundreds of thousands of pounds per deployment and exists in fewer than 100 stadiums globally. The other 99 percent of cricket has nothing.
Our Approach
Howzzat.ai is building DRS-equivalent decisions using two ordinary Android smartphones and computer vision, instead of specialist hardware. Two phones positioned at square leg and behind the bowler capture video at 1080p 60fps. A YOLOv8 object detection model locates the ball frame by frame, and a centroid tracker reconstructs its trajectory. For wide decisions, that trajectory is compared against calibrated crease positions, set through a one-time four-point calibration that adapts to any ground without fixed hardware. The next milestone, LBW detection, combines trajectory data from both cameras to model the ball’s path in 3D relative to the stumps.
The hard part is the gap between stadium conditions and a real ground: a small, fast ball, consumer lenses, inconsistent lighting and camera placement, and no fixed, synchronised, high-speed cameras to rely on. Printmind is Howzzat.ai’s technical build and test partner, contributing the AI product strategy and computer vision engineering behind the tracking, calibration and decision pipeline.
The Impact
The core pipeline (ball tracking, crease calibration, decision classification and annotated video output) is built and runs end to end on test footage. Live field testing starts this month at club cricket in London and at a turf ground in Hyderabad, targeting 85 percent plus precision on wide detection under real match conditions before any ground deployment.
A savings marketplace that keeps every pound inside FSCS protection, automatically
Zavings (zavin.gs), a Printmind venture.
The Opportunity
UK savers hold an estimated £276 billion in accounts earning no interest, according to the Bank of England, while others unknowingly hold balances above the £120,000 FSCS protection limit at a single bank. Chasing a better rate elsewhere means a fresh application, a new login and another identity check, so most people never move their money. On the other side of the market, banks that want to attract deposits face expensive acquisition costs to win savers one at a time.
Our Approach
Printmind is the design and build partner behind Zavings, a two-sided marketplace connecting savers directly to partner banks. The work has covered the full concept: the Zavings brand and identity, an end-to-end clickable prototype for both the saver app (a single ID check, a funded wallet, live rate comparison, FSCS-aware allocation across multiple banks, and withdrawal) and a partner-bank console (list products, go live, receive pre-verified deposits, settlement), and the go-to-market groundwork, including a waitlist landing page, an investor one-pager, and an analysis of the regulatory authorisation routes.
Printmind also shaped the underlying model: the FSCS-aware allocation logic that automatically keeps a saver’s cash under the protected limit at each bank, and the e-money regulatory path under which partner banks remain the deposit-taker of record.
The Impact
Zavings is a pre-seed concept with a working end-to-end prototype. It is not yet live and not yet regulated to hold customer money. Validation so far is a single, encouraging conversation with one partner bank, confirming the problem on both sides of the market rather than proving demand at scale. The current work is broadening that: further partner-bank conversations, a public waitlist to size retail interest, and a paid regulatory scoping session to confirm the authorisation route.
Putting the patient at the centre of India’s prescription journey
RxSpeed by Health OK!, a Printmind venture.
The Opportunity
In India, the journey from a doctor’s prescription to a filled prescription is fragmented, and the patient is nobody’s customer. Prescriptions are often handwritten, pharmacies dispense with no verification loop back to the doctor, and patients are left with no record of what they have taken, what remains on a course, or what it should cost. The major platforms in the space, including Eka Care, HealthPlix, Netmeds and PharmEasy, are each built doctor-first, hospital-first or pharmacy-first. None puts the patient at the centre of the loop.
Our Approach
RxSpeed is the first product from Health OK!, a digital health venture focused on patient-first infrastructure for Indian healthcare. Printmind is RxSpeed’s design and build partner and technical delivery engine. The prototype, a mobile-responsive application covering the full doctor to patient to pharmacy flow, was designed and built through Printmind’s studio capability: a sub-20-second digital prescription for the doctor, built-in safety checks for allergies and drug interactions, and for the patient a portable prescription history, pharmacy price comparison, and the ability to split dispensing across pharmacies.
The platform is being architected for AI participation from day one, with structured, machine-readable prescription data and ABDM and ABHA interoperability treated as first-class design constraints rather than added later.
The Impact
RxSpeed has a functional, hosted prototype and a pilot design with three sequenced go or no-go gates: doctor retention, prescription capture rate, and pharmacy willingness to grant a platform discount for verified demand. The pilot location is Sainikpuri, Hyderabad. Field testing has not started yet. What exists today is qualitative validation in progress: clinical review of the prototype by an NHS GP, and pharmacy identification underway in Sainikpuri ahead of the first gate.
Working on something with a similar shape?
If you are reading the whole business before proposing a fix, or building against measurable KPIs rather than a fixed spec, we probably think about the problem the same way you do.