Eleven plays we run on every engagement

1. 01

   Play 01

   Start with what your customers need

   Projects fail when they're built around an org chart instead of the people using them. Before we design or automate anything, we find out what your customers and your team are actually trying to do.

   • Talk to the people who will use the thing — customers and staff
   • Watch how the work happens today, including the workarounds
   • Write down the tasks people are trying to complete, in their words
   • Test early versions with real users, not just stakeholders
2. 02

   Play 02

   Design the whole journey

   A fast website feeding a broken follow-up process is still a broken experience. We map everything from the first click to the invoice getting paid, and fix the handoffs — not just the homepage.

   • Map every step, system, and person a customer touches
   • Find where things get dropped, retyped, or forgotten
   • Design the online and offline parts as one process
   • Measure the journey end to end, not screen by screen
3. 03

   Play 03

   Keep it simple and obvious

   Every extra field, step, or clever label costs you customers. We use plain language, cut steps that don't earn their place, and make the next action obvious on every screen.

   • Use the words your customers use, not industry jargon
   • Remove every form field that isn't strictly needed
   • Make one clear next action per page
   • Design for the phone in someone's hand, not the designer's monitor
4. 04

   Play 04

   Ship early, then iterate

   A working slice in three weeks beats a big reveal in three months. We get something real in front of you fast, then improve it in short cycles based on what actually happens.

   • Deliver a usable version within the first weeks, not at the end
   • Ship improvements in small, frequent releases
   • Let real usage — not opinions — set the priority list
   • Keep a public roadmap of what's next and why
5. 05

   Play 05

   Automate the boring parts

   If a task happens twice a week and follows rules, it belongs in a workflow. We look for the copy-paste jobs, the re-keyed data, the Friday reports — and hand them to software.

   • Inventory the recurring tasks and what they cost in hours
   • Automate the rule-following work first; it pays back fastest
   • Use AI for the reading, drafting, and sorting steps
   • Leave people doing the work that actually needs a person
6. 06

   Play 06

   Keep a human in the loop where it counts

   Automation should draft, sort, and prepare — and pause for a person wherever money moves, promises are made, or judgment matters. You stay in control of the decisions that are yours.

   • Decide explicitly which steps need human approval
   • Make approvals one click, with full context attached
   • Log every automated action so there's always an audit trail
   • Design escalation paths for anything unusual
7. 07

   Play 07

   Choose boring, modern technology

   We build on mainstream, well-supported tools — Next.js, n8n, Zapier, managed cloud services — so what we hand over is something any competent developer can pick up, not a science project only we understand.

   • Prefer widely adopted tools with real communities
   • Use managed services instead of servers to babysit
   • Avoid anything that locks you to us or to a vendor
   • Pick the simplest stack that genuinely fits the job
8. 08

   Play 08

   Automate testing and deploys

   Changes should ship through a pipeline, not a late-night FTP session. Every change is tested and deployed automatically, with a preview link you can click before it goes live.

   • Every change gets a preview environment to review
   • Tests run automatically before anything reaches production
   • Deploys are one click — and so are rollbacks
   • Workflows are versioned and tested like code, because they are
9. 09

   Play 09

   Build in security and privacy

   Automation touches your CRM, your inbox, your billing. Every connection we build uses least-privilege access, secrets stay out of code, and customer data goes only where it needs to.

   • Scope every API key to the minimum it needs
   • Keep credentials in secret managers, never in code or Zaps
   • Know what data flows where, and write it down
   • Review access when the project ends, not just when it starts
10. 10

    Play 10

    Measure what matters

    Every site ships with analytics and every workflow ships with run logs. Decisions about what to build next come from what the data shows, not from whoever spoke last.

    • Define what success looks like before building
    • Instrument sites with analytics and workflows with run histories
    • Review the numbers together on a regular cadence
    • Kill or fix what the data says isn't working
11. 11

    Play 11

    Hand over everything

    Your code, your accounts, your data, your documentation. When an engagement ends you own the whole system and understand how it works — hiring us again should be a choice, never a hostage situation.

    • Everything lives in accounts you own from day one
    • Code, workflows, and credentials are documented as we go
    • Handover includes a walkthrough, not just a zip file
    • Anything we'd need to maintain it, you have too

Adapted for private-sector client work from the U.S. Digital Services Playbook, published in the public domain by the United States Digital Service.