APPROACH
Deep experience.
Close to your business.
We choose work where we can understand the whole operation and stay close to the people using what we build.
Experience, judgment and taste turn that understanding into a coherent design: what to keep, what to connect and what is worth building.
THE EXPERIENCE BEHIND THE WORK
From the plant floor to the media plan.
A view of the whole operation.
Our practice draws on 20 years of experience across industrial automation, intraplant logistics and software integrations.
This experience comes from earlier roles and projects across industrial operations, digital advertising technology, internal tools and customer-facing systems. It now informs the advice, design and delivery available through Codewrights.
Industrial operations
Deep experience serving the commercial laundry industry, end to end: operations, controls, data, internal tools and customer-facing software.
That work includes technical design for new installations, updates to existing hardware controls, software around installed equipment, and new operations designed around software capabilities.
Digital audiences and telemetry
At Centro, now Basis Technologies, Naveed Kakal led and managed the teams that built many of the platform’s original DSP integrations and its media planning tools. That meant learning digital audience targeting and telemetry collection from the inside: what gets measured, how it moves between systems and what it can honestly say.
We bring the most useful parts of that world to your business, sized for a boutique footprint: first-party measurement of where visitors come from, campaign tags that carry through to your own reports, and audiences you build with consent, without renting an ad-tech stack.
Working across those disciplines, both internally and for customers, shapes how we approach a business: understand the whole operation, connect the parts, and carry the work from design through delivery and ongoing care.
APPLIED AI EXPERIENCE
Knowledge that stays.
Work that moves.
Our experience also includes building AI into the work itself: helping teams investigate problems, turn incoming messages into action and put shared knowledge to use.
These capabilities come from earlier software projects. We can adapt them into a focused tool or a connected part of your existing systems, shaped around your team’s information, decisions and ways of working.
AI for industrial operations ↗AI-assisted diagnostics
Bring reported symptoms, photos, equipment context and technical documents together to suggest likely causes and next checks. Give the team a useful starting point for investigation, with the evidence alongside it.
Omnichannel intake & workflow routing
Recognize what an incoming email or text is asking for, extract the relevant details and route it into the right workflow. Prepare an order change or follow-up for staff review, keeping the original message and customer context attached.
Team knowledge, where the team talks
Capture useful know-how from team conversations and make it available to answer later questions inside the chat tools people already use. Connect answers to the underlying conversations and documents, so hard-won knowledge is easier to find and reuse.
CRAFT AND TOOLS
More room for the work that matters.
Modern tools, including AI, help us explore, build and refine. They make a depth of design and delivery that once needed a much larger team practical at a smaller scale.
We stay responsible for the architecture, the details and what ships. Features earn their place by helping the business, and the systems you already depend on are part of the starting point.
HOW AN ENGAGEMENT RUNS
From how the work happens to keeping it running.
We work forward-deployed: on contract, close to your team and the decisions that shape the work. Collaboration can be remote. You talk directly to the people advising, designing and building, with fewer handoffs and less context to repeat.
Understand the business
We start with how the work really happens: who does what, what customers hear and when, and where it goes wrong today.
Choose the useful first release
Not everything at once. The first release is the smallest one that changes a working day.
Build around what works
We identify the systems to keep, the connections to make and the gaps worth building. Your existing software and corporate data are part of the design.
Handle the edge cases
A failed payment, a customer who isn’t home, a reply nobody expected. Each gets an owner and an outcome, never a silent failure.
Know what is ready to ship
Each workflow has a written definition of done. It becomes tests before the workflow counts as finished.
Keep it running
Releases, monitoring, backups and changes stay with us after launch.
WHO DECIDES WHAT
Clear responsibilities.
A shared understanding.
Yours
Business decisions and commitments: what you sell, what you promise customers, what you spend.
Ours
Technical judgment, implementation and operation within the engagement, so you are not coordinating a software stack.
RULES FROM A REAL BUILD PLAN
What “done properly” looks like.
These are the delivery rules from the pest-control case study, unedited. Every plan gets its own.
- Company facts, plans, clauses, templates, rules and brand tokens are data. No company literals in code.
- Money is integer cents, and every money row carries the processor’s id.
- Every state change writes an event in the same transaction. Timelines and automations read events; they never infer history from current state.
- Nothing customer-facing sends without a rule or a person, consent for that purpose, quiet hours and a dedup key.
- Jobs are idempotent and safe to retry. Webhooks are verified.
- Every message carries a signed, expiring, single-purpose link. Signing in is the exception.
- Roles are checked on the server for every office, technician and rep route.
- Phone-first for technicians, reps and customers; dense desktop for the office.
- Accounts and properties are many-to-many through grants. Every read of a property, visit, report or invoice checks the grant, not just the account.
- Boring infrastructure: Rails, one Postgres database, Solid Queue, one Fly app per environment.
- Every tenant table carries a company id, and a test proves one company can never read another’s rows.
- Each workflow’s acceptance list becomes system tests before the workflow counts as done.
START WITH A CONVERSATION
Tell us what the business runs on today.
The tools, the workarounds, the things that fall through. We’ll tell you what we would build, what we would keep and what we would leave alone.
Start a conversationinfo@codewrights.io