A software development company that stays after launch
Senior engineers design, build and maintain the system. Twenty-two of them are in production right now.
Hashlogics has 22 custom systems running in production, and the engineers who built them are still the ones who fix them. That is the whole difference. A build that reaches launch is only halfway done. Expensive failures arrive later. An integration quietly drops rows when a partner changes their API. An access rule leaks one tenant's data into another's report.
Why the second year is where software fails
You can get a working demo from almost anyone now. Getting one that survives real users is a different job. A demo holds ten records and one happy path. Your business holds ten years of messy data and four systems that must agree. Then someone does the unpredictable thing on a Friday afternoon.
Most of what we are asked to fix was shipped by someone competent. It broke because nobody owned it afterwards. We wrote separately about surviving real users, and this is most of what that means.
- 01The vendor stops answering once the last invoice clears.
- 02Nobody left can read the code, so every small change gets quoted as a rebuild.
- 03The engineer who understood the data model has moved on.
- 04A partner changes an API and nothing tells you until the numbers stop matching.
What is running now
Counted, not estimated
22
custom systems in production
70%
less manual compliance checking at Shift Link
5
permission tiers governing TankAware
8
builds taking payments through Stripe
Three shapes this work takes
The system that runs the business
Operations platforms, compliance tooling, internal products that replace a spreadsheet nobody trusts. WAIQ standardises procedures and training across sites, departments and roles in one place.
See WAIQ →
A product you sell to other people
Multi-tenant SaaS with billing, roles and an audit trail. Eight of our builds take payments through Stripe, so the ledger questions are ones we have already answered.
How we store money →
The integration nobody wants to own
Connecting an ERP, a government API or a warehouse to a workflow that has to keep agreeing with them. Shift Link checks right-to-work status against a UK Government API before it will schedule anyone.
See Shift Link →
The AI feature that has to be right
Agents and retrieval built into the product, with an evaluation gate rather than a vibe check. Trading Copilot asks a human to approve every trade before it executes one.
Prototype to production →
- ScopeFree call. We map the workflow.
- DiagnosePaid, only inside existing code.
- ArchitectData model settled before code.
- BuildSenior engineers, live environment.
- RunService level, or trained handover.
The diagnostic is paid because reading someone else's codebase is real work. Scoping a new build is not, so that call is free.
Architecture is decided before anyone writes code
Most rewrites we are called into trace back to one early decision. A data model that could not hold a second currency. An access rule written in the API instead of the database. Money stored in a float.
None of those hurt in month one. All of them are expensive in month eighteen. So we settle them first, on a call where you can argue with the answer.
- Access rules live next to the data, not in whichever route remembered to check.
- Money is stored as integers or a fixed-precision decimal, never a float.
- Anything a retry could duplicate carries an idempotency key.
What we actually ship on
Application
Data and platform
In the product
Named clients, running systems
TankAware
AI + IoT platform for petroleum site management.
Read the case study →
Maidily
Integrated operations platform for residential cleaning businesses.
Read the case study →
WAIQ
Business operations and execution platform for multi-site organizations.
Read the case study →
Elevent
Real-time multiplayer trivia for corporate events and training.
Read the case study →
“Their attention to detail, quality of employees, and work ethic were outstanding.”
Nicolas de Quesada · CEO, Lexpair
The usual engagement against ours
Who writes the code
The usual approach
A senior engineer wins the pitch. Juniors do the build.
How we run it
The engineers on the scoping call are the ones who ship it.
After launch
The usual approach
Support is a separate contract, negotiated once something breaks.
How we run it
Every engagement ends with a decision made up front: an agreed service level, or a documented handover with training.
Pricing
The usual approach
An hourly meter, or a number quoted before anyone has seen the problem.
How we run it
Fixed after scoping. We are not the cheapest bid and do not try to be.
The AI part
The usual approach
A chatbot bolted onto the side of the product.
How we run it
Agents and retrieval built into the workflow, with an eval gate before release.
What you own
The usual approach
A deployment you cannot rebuild without the vendor.
How we run it
The code, the data model and the infrastructure, documented so your team can read it.
What buyers ask first
01What happens to the software after you launch it?+
Whoever built it keeps running it, under an agreed service level, or your team takes it over with documentation and training. That choice is made before the build starts, not after something breaks. The first 2 months of support and maintenance are free, with every build. This is the question we get asked last and it should be asked first, because it is where most custom software quietly dies.
02How much does custom software development cost?+
Four things drive the cost. How many systems it must integrate with. The state of your existing data. Whether a security review stands between you and launch. Who owns the thing afterwards. Integration count moves the number most. We publish no price ranges, because a figure quoted before anyone has read your workflow is a guess wearing a decimal point. You get a fixed price after scoping.
03Should we build custom software or buy something off the shelf?+
Buy it if the process is one everybody runs the same way, such as payroll or email. Build it when the process is how you compete, or when no product on the market will agree with the three systems you already run. Our build vs buy software page walks the decision with a table you can take into a meeting.
04Do you work with startups or only enterprises?+
Both, at the same standard. An enterprise build starts from an existing stack and a security review. A startup build starts from nothing and has to convince an investor's engineers. Neither gets a discounted bar. Early shortcuts are the most expensive thing a young product can carry.
05Do you charge for the first call?+
No. Scoping calls are free. A paid two-week diagnostic applies in one case only. That is where we go into an existing codebase to work out what is really in there. Reading someone else's system is engineering, not a sales conversation.
06Can you build the web app and the mobile app?+
Yes, and most of our systems are both. Shift Link runs a Next.js web application for schedulers and a React Native app for workers in the field. Our web application development and mobile app development pages cover each side in detail.
Go deeper
- Web app development →The product your customers log into.
- Mobile app development →Apps that still work with no signal.
- Build vs buy software →The decision, with a table you can take into a meeting.
- From prototype to production →What changes when the demo meets real users.
- Best custom software development companies →A ranked shortlist, with the methodology stated.
