Solutions
Eight services. One team that has to make them work together.
You can hire us for one of these. Most clients start that way. The reason they expand is that the problems turn out to be connected — the site cannot convert what the ads send if the CRM loses it, and the portal cannot report on data the backend never captured.
01 · Engineering
Website & Web Platforms
Corporate sites, product sites and web applications that load fast, rank well, and can be edited by the people who own the content.
Most sites get slow and unmaintainable for the same reason: every page is built as a one-off. We start with a component library and a content model, so the marketing team assembles new pages from tested blocks instead of filing a ticket with us every time.
That matters most for companies with a real product range. If you have fourteen products and each one needs its own specification depth, hand-building fourteen templates is how a site becomes a liability within two years.
- Component library and design system documented for your team
- Headless or traditional CMS, mapped to how your team actually writes
- Core Web Vitals budget agreed up front and tested on every release
- Accessibility to WCAG 2.2 AA, checked with real assistive technology
- Structured data so search and answer engines can parse your pages
- Analytics and conversion tracking wired in before launch, not after
02 · Engineering
Customer Portals & Custom ERP
The systems your staff, dealers and customers log into every day — replacing the spreadsheet and WhatsApp group that currently hold the business together.
Off-the-shelf ERP fails when your process does not match the vendor's assumptions. You end up paying for modules you do not use and working around the ones you do. Custom is the right answer when your process is genuinely your competitive advantage — and the wrong answer when you are just avoiding change management.
We will tell you which situation you are in. Sometimes the honest recommendation is to configure something existing and spend the budget elsewhere.
- Role-based access control with a real permission model, not just hidden buttons
- Approval and workflow engines that match your escalation reality
- Audit logs on every state change — who did what, when, and what it was before
- Integrations with the systems you already run: accounting, CRM, telematics, email
- Reporting views that replace the manual monthly spreadsheet assembly
- Bulk import tooling, because your data migration is never as clean as promised
03 · Engineering
Backend, DevOps & Infrastructure
APIs, data models, pipelines and the cloud plumbing underneath — built so the 3am page is rare and diagnosable when it happens.
Uptime is a design decision, not a monitoring product you buy afterwards. We handle the parts that get skipped under deadline pressure: migration strategy, backup restore testing, rate limiting, secret rotation, and observability that tells you which of the twelve services is actually the problem.
If you inherited a system nobody fully understands, we do audits too — a written read on what is fragile, what is fine, and what to fix first.
- API design with versioning strategy, so clients do not break on your releases
- CI/CD pipelines with automated tests and one-command rollback
- Infrastructure as code — reproducible environments, no snowflake servers
- Monitoring, structured logging and alerting that distinguishes urgent from noisy
- Backup and restore procedures that have actually been rehearsed
- Security review: dependency audit, secret handling, access boundaries
04 · Engineering
AI Infrastructure
AI built as infrastructure with retrieval, evaluation and guardrails — not a chatbot bolted onto the homepage because the board asked about AI.
The gap between an impressive demo and a system you can put in front of customers is almost entirely evaluation and failure handling. A model that is right 85% of the time is a support liability unless you know which 15% it gets wrong and what happens then.
We build the unglamorous parts: retrieval over your actual documents, evaluation sets that catch regressions, human review where stakes are high, and cost controls so the bill does not surprise you in month three. We run these systems internally before we recommend them — see the AI Lab.
- Retrieval over your documents and data, with citations back to the source
- Evaluation harness so you can measure whether a change actually improved things
- Guardrails, fallbacks and human-in-the-loop review on high-stakes paths
- Cost and rate controls with per-feature budget visibility
- Prompt and model version control — reproducible, reviewable, rollback-able
- Honest scoping: we will tell you when a rules engine beats a model
05 · Demand
SEO & AEO
Technical SEO plus answer-engine optimisation — being found when the buyer asks a search engine, and when they ask an AI assistant.
Search is splitting in two. Traditional rankings still drive volume, but a growing share of research now happens inside AI assistants that synthesise an answer without a click. Being cited in that answer requires different work: clear factual structure, retrievable claims, and content that reads as a credible primary source.
We do both, and we report on both. No keyword-stuffed volume plays — that approach is actively penalised now and it damages the brand you are building.
- Technical audit: crawl, index coverage, Core Web Vitals, structured data
- Content architecture built around real buyer questions, not keyword volume
- AEO work: entity clarity, citable structure, factual precision
- Internal linking and topical authority built deliberately
- Reporting against pipeline and qualified enquiries, not just position
- Editorial standards — every piece has a named human editor accountable for it
06 · Demand
Paid Media
Search, social and retargeting managed against cost per qualified opportunity — the number that affects the business, not the one that flatters the dashboard.
Most paid accounts we inherit are optimising toward form fills, with no visibility into which of those fills became revenue. That means budget flows toward whatever generates the most junk. Fixing the measurement usually improves results more than fixing the creative.
Because we typically also built the site and the CRM, we can close that loop properly — tracing a closed deal back to the campaign that started it.
- Account structure rebuilt around margin, not just volume
- Conversion tracking that reaches into the CRM, including offline closes
- Creative and landing page testing with enough traffic to be conclusive
- Negative keyword and audience hygiene, reviewed continuously
- Budget pacing with a clear read on incremental versus cannibalised demand
- Monthly reporting a CFO can follow, not a screenshot of the ads dashboard
07 · Demand
Digital Marketing & Lifecycle
Content, email, lifecycle and analytics — the connective work that turns traffic into pipeline and pipeline into repeat business.
Acquisition gets the attention and the budget. But for most businesses the cheaper win is in the middle: the leads that went quiet, the trial users who never activated, the customers who would buy the second product if anybody had told them it existed.
We build those sequences and measure them against revenue, so you can see which ones deserve more investment.
- Content strategy tied to the actual buying committee and their objections
- Email and lifecycle automation with segmentation that reflects behaviour
- Analytics implementation you can trust — one number, one definition
- Attribution modelling appropriate to your sales cycle length
- Sales enablement material that the sales team actually uses
- Quarterly strategy review with a written recommendation, not a slide deck
08 · Operations
Support, Training & Enablement
Technical support desks, team training and documentation — so the capability stays with your team after we step back.
A system nobody was trained on gets abandoned within a quarter, regardless of how well it was built. We treat handover as part of the build, not a courtesy at the end.
For some clients we also run the support function directly — L1 and L2 technical support calling, ticket triage and escalation — while their internal team is being hired and trained up.
- L1/L2 technical support desks with defined escalation paths
- Live training sessions plus recorded walkthroughs your team can revisit
- Written documentation aimed at the next engineer, not at us
- Runbooks for the operational tasks that used to live in one person's head
- Onboarding material for new hires joining an existing system
- A deliberate exit plan — we build so you are never trapped with us
Engagement models
Three ways to work with us
Whichever model you pick, you own the code and the accounts, and you get a named engineer plus a named account lead.
Fixed-scope project
Defined deliverable, written statement of work, fixed price. Best when you know what you need — a site rebuild, a portal, a migration. Changes go through a written change note so the invoice never surprises you.
Monthly retainer
An agreed capacity per month across engineering, marketing or both. Best for continuous improvement after launch, or when priorities shift faster than a fixed scope can accommodate. Month-to-month, no annual lock-in.
Embedded team
Dedicated engineers and specialists working inside your process, your standups and your repositories. Best when you are scaling faster than you can hire and need capability now rather than in six months.
Next step
Tell us what is actually broken.
A 30-minute call with an engineer, not a salesperson. You will get a straight read on whether we are the right team — including if the answer is no.