
A2A Protocol Explained
Google's A2A protocol lets any AI agent talk to any other AI agent. What it does, how it works, and why it changes everything for businesses.
Long-form, practical guides on scoping SaaS MVPs, building custom software, integrations, customer portals, dashboards and handover. Written for founders and product teams who want to ship owned software without guesswork.
56 articles

Google's A2A protocol lets any AI agent talk to any other AI agent. What it does, how it works, and why it changes everything for businesses.
Pricing systems need plan logic, approval, version history and buyer behaviour, not another duplicated doc.
Live reporting works when the data model is designed around how the client reads performance.
When the same handover breaks every week, software may be the cleaner hire.
Fix time and budget, then scope. One workflow, one system, clear handover.
Send proposals while the buyer still remembers the call. Template logic beats copy-paste.
Stop spreading updates across email, support tools, Notion and Drive. One owned place for status, files and approvals.
Pull product, revenue and CRM data into one view instead of rebuilding reports.
Route leads, track delivery and surface blockers in one workspace, not five tabs.
100% code ownership, deployment notes and workflow logic stay with you. No licence lock-in.
Run cleaner when project status lives in owned software your team trusts.
Use SaaS when the problem is solved. Build when your workflow is the differentiator.
Collect briefs, assets and approvals in one auditable flow, not twelve threads.
SaaS builds from £5,000 and Enterprise from £18,000, with fixed scope and full handover.
Pricing, reporting, onboarding, customer comms and glue work, mapped to what to build first.
Source code, deployment access, docs and clear ownership. No mystery system.
One view for founders: pipeline, workload, margin and what needs attention today.
Reusable sections, pricing blocks and sign-off, without rebuilding from scratch.
Product analytics, Stripe, HubSpot and CRM, connect the data you already have.

Scoped around one bottleneck: onboarding, reporting, portals or internal ops.
You work directly with the person designing and shipping, not a handoff chain.
Clear deliverables, assumptions and acceptance criteria beat vague promises and over-lawyered walls of text.
Same steps before every build: brief, access, stakeholders, success metric and internal owner.
Headline, what moved, why it moved, next actions, structure commentary so updates stop taking all day.
Merge duplicates, close zombie deals and enforce owners before any system downstream can be trusted.
Done, next, blocked, risks, a written rhythm clients trust without another calendar block.
Your pricing system should enforce how you sell, not let reps accidentally mix models.
Draw the chain, mark wait times, circle the longest gap, fix policy before you buy tools.
Score fit, repetition, integration, ownership and glue work, same sheet every time you evaluate.
Mutual response times, revision rounds and what happens when feedback is late, agreed in onboarding.
Property structure, key events, UTMs and QA, productise the checklist instead of relying on memory.
SOW exclusions, visible request logs and weekly out-of-scope review, process before personality.
Signed scope, call summary, margin notes and comms preferences, success accepts or rejects, not inherits.
Rank tool pairs by hours saved and data quality, one direction, one object type for version one.
If the meeting produces no decision, cancel it, replace readouts with async updates.
Track at project level, compare sold vs actual monthly, feed learnings back into pricing.
Problem, users, integrations, success metric, budget and ownership, enough to compare vendors fairly.
Structured data, allowed actions, human approval and audit logs, one workflow at a time.
Monolith, auth, database and API boundaries for a first SaaS product that can ship in weeks, not months.
Subscriptions, usage limits and webhook handling without over-engineering payments on day one.
Pick auth that matches your buyers, not your favourite library. Sessions, invites and admin access.
When a standard Postgres stack beats Firebase, and what to optimise for in the first 90 days.
Ship behind a private link, promote to production safely, and hand over access the client can run.
Microservices add coordination cost before you have product-market fit. Start simple, split later.
Admin, member, viewer and customer-facing roles. Model what buyers actually ask for in sales calls.
Expose what integrations need, hide what they do not. Stable IDs, pagination and error shapes.
Speed with guardrails. Copy-paste is fine in UI; data model shortcuts are not.
Time-to-value, activation events and support tickets. Measure the funnel before you redesign UI.
How to model accounts, teams and billing without building enterprise complexity on day one.
When Looker or Metabase is enough, and when operators need a dashboard tied to your workflow.
Users, data sources, approvals, integrations and success metrics. A worksheet for the first call.
A simple model to justify custom software to finance when spreadsheets are still 'free'.
Practical steps to define scope, avoid creep, and get from idea to launch in weeks, not months.
Timelines, communication, and how Standen delivers scoped SaaS and custom software builds for founders.
Deciding between bespoke build and off-the-shelf tools? We break down the trade-offs so you can choose with confidence.
How we ship MVPs and web apps in weeks while keeping quality high and tech debt low.
No articles match your search.
Book a short call. We’ll map the simplest system worth building first.
Book a call