Every page on the site, in one list.
Every service, platform, case study and article. Filter them by platform, technology or the problem you are trying to solve.
Services
The disciplines: process automation, AI, custom development, integration.
Platforms
The platforms we build on, one page each.
Solutions
What we build for finance, operations and integration problems.
Integrations
How specific pairs of platforms connect, and where they break.
Markets
UK, US, Australia and Pakistan: what differs in each market.
Work
Case studies from client engagements and Skofi's own products.
Insights
Comparisons and engineering notes on automation platforms and AI.
About
Who Skofi is, how the work runs, and who is accountable for it.
Contact
Book a 15-minute call or send an email.
Press and media kit
Facts, founder bio, descriptions and logos for anyone writing about Skofi.
Custom Software Development
When the SaaS runs out, build the thing that fits the business.
AI Engineering & Integration
Put models where they do measurable work, and keep a human on the decisions that matter.
Claude Code Development & Team Setup
Software built with Claude Code, with an engineer reading every change.
MCP Server Development
Give AI agents controlled access to your systems, with permissions and a log.
Vibe-Coded App Rescue & Production Hardening
You prototyped it with AI. We make it safe to put real customers on.
Business Process Automation
Take the manual steps out of the work that actually runs the company.
Systems Integration & APIs
Make the systems you already pay for behave as one.
Workflow Automation
Make.com, Zapier and n8n scenarios built to survive contact with production.
Ops Platforms & CRM
monday.com, Airtable and HubSpot set up so the team actually uses them.
Make.com
Scenario design, rescue and cost work on Make.com: routers, iterators, aggregators, error handlers, and blueprints you can commit without leaking a token.
monday.com
monday.com automation and API work: boards modelled around what an item really is, statuses that carry one meaning, and portals instead of seats.
Airtable
Airtable schema design, linked records and rollups, interfaces and automations, plus an honest read on where a base stops and a database should start.
HubSpot
HubSpot integration and workflow work: property architecture, enrolment that behaves, and syncs to the ops systems where delivery actually happens.
Zapier
Zapier build, audit and migration work: Zaps that stay maintainable, task costs kept down, and an honest view of when the logic has outgrown the platform.
n8n
n8n workflow build and self-hosting: nodes, credentials and executions done properly, with the upgrade, backup and encryption-key work nobody budgets for.
Claude
Claude API integration, Claude Code setup for teams and AI agents for business operations, with review steps, evals, cost caps and an audit trail.
Custom ERP development
Custom ERP development for small and mid-sized firms: modules, permissions and audit trail on Postgres, and when buying is the better answer.
AI invoice processing
AI invoice processing that reads supplier invoices, validates the output and waits for a person to confirm before anything reaches the ledger.
Systems & API integration
Systems and API integration consultant: syncs, webhooks and custom connectors with idempotency, a decided system of record and a run log you can read.
Finance workflow automation
Workflow automation for finance teams: approvals that gate the totals, one ledger so reports reconcile, and a plain account of what we have not built.
Solar operations software
Operations software for commercial solar installers: deals, sites, surveys, consents and delivery in one system, with one project opened per building.
Claude API integration
Claude API integration consultant: extraction, classification, drafting and tool use with review before write, failover, evals and spend limits.
Internal tools development
Internal tool, admin panel or portal development: roles, audit trail, soft delete and exports, and when a platform is enough. Owned by you from day one.
Vibe code security review
Vibe code security review: a ranked written report on an AI-built app's access control, secrets, database rules and spend caps. Not a penetration test.
monday.com + Make
Connect monday.com to Make: instant versus polling triggers, column-value shapes, rate limits, duplicate items, and when a board needs a server instead.
HubSpot + Airtable
Sync HubSpot and Airtable without duplicates or loops: field ownership, ID mapping, merged records, rate limits, and four ways to build it.
n8n + HubSpot
Connect n8n to HubSpot: webhook triggers, signature checks, deduping retries, rate-limit batching, and when HubSpot workflows should do the job instead.
Zapier to Make
Migrate Zapier to Make safely: no Zap converter, how task and credit costs differ, a parallel-run cutover, and which Zaps should stay where they are.
United Kingdom
A UK finance or operations team automating work runs into three things a generic build ignores: VAT records that have to stay digital end to end, accounting software with its own API habits, and UK GDPR duties over where personal data travels.
United States
A US team automating operations meets four things early: QuickBooks as a widely used ledger, sales tax that varies by state and sometimes city, a CRM that is often HubSpot or Salesforce, and a security questionnaire that can arrive mid-deal.
Australia
Australian operations builds are shaped by GST and the BAS cycle, payroll that reports to the ATO on every pay run, Xero or MYOB as the ledger, the Privacy Act's rules on offshore data, and a calendar of public holidays that differs by state.
Pakistan
Two different readers arrive at this page. One runs a business in Pakistan and has local tax, invoicing and payment constraints to design around. The other is a Pakistani freelancer or firm whose clients are overseas and who found them through a marketplace such as Upwork.
An ERP with AI invoice capture and an audit trail that holds
Eighteen modules replacing a legacy desktop ledger, with supplier invoices captured by photograph, approvals that gate revenue, and a field-level audit log nobody can rewrite.
Orders that file themselves, and an SLA clock that never forgets
Inbound orders parsed straight onto monday.com boards, installers allocated and notified automatically, and a two-business-day SLA engine with escalation, shipped behind a shadow-mode flag.
A marketplace that recruits whichever side of itself is short
A public website and a signed-in marketplace in one codebase, with a home page that reads live supply to decide which side to argue for, prices held in the database, and contact details the server withholds until they are asked for.
An ops platform where the building is the record, not the deal
CRM, site records, surveys, quotes, consents and delivery in one system, where a deal covers as many buildings as the negotiation does, and winning it opens one project per building.
A staffing platform that keeps its rules in the database
Recruitment, right-to-work, rostering, check-in, pay and client paperwork for a large pool of zero-hours workers: three front ends over one Postgres database that refuses to break its own rules. Built and tested on staging, awaiting the client's content and keys.
Heat-pump quotes with the compliance gates written into the code
From website enquiry to a fixed-price quote with no site visit: a regulated sales platform where the noise assessment, the designer's sign-off and the statutory cooling-off period are gates the system enforces. In build.
One login for a solar business, in place of a patchwork of tools
CRM, projects, quotes, purchase orders, outreach, a customer portal and an AI copilot in one connected system, with a pipeline that refuses to advance a deal until the data to deliver it exists, and a team-test loop that has shaped it through sixty-one tickets.
A content engine that is not allowed to make things up
An AI-assisted marketing operations platform (weekly LinkedIn drafts, sourced news, funding alerts, competitor tracking and a media library) where funding and statistics claims need a verified source and nothing publishes without approval.
An estimating assistant where the model names the parts and code sets the prices
Past proposals become a searchable library, and a plain-English brief becomes a priced draft estimate in which every line says where its number came from, built as a working demo on the integrator's own past proposals.
Blueprint Runner: a self-hosted engine that speaks Make natively
Paste a Make.com blueprint export, keep the tokens in variables rather than the JSON, and run the scenario on your own infrastructure, with a per-module log of every request and output.
Skofi Outreach: deliverability rules the database refuses to break
A prospecting and email-sequencing platform where the three rules that protect your sending domain are Postgres constraints, not application logic.
FlowForge: a workflow builder with cron, webhooks and run history
Multi-flow visual workflow builder deployed as serverless functions: per-flow schedules, public webhook triggers, background execution and a server-side secrets vault.
FormCraft: a form platform with an API automation can call
A form builder replacing a per-seat SaaS product, with a token-secured REST API and an outbound webhook so Make can both read from it and react to it instantly.
School management ERP with permissions enforced per branch
Students, fees, payments, payroll and accounting across multiple branches, with role and branch permissions enforced in the database and fee reminders sent over WhatsApp.
A privacy-first health PWA with consent-scoped sharing
Deep domain modelling, an import path off a competitor's export format, and partner sharing where every category is a separate consent enforced in the database.
Make vs Zapier vs n8n: choose by billing unit
Zapier bills per task, Make in credits, n8n Cloud per execution. How to count your own units, and when each platform is the right choice.
Airtable vs monday.com: pick by your core object
Airtable suits records that relate to each other; monday.com suits work that moves through stages with owners. Decide by your core object.
Build or buy: the three numbers that decide it
Buy by default. Build only when licence drift, integration tax and workaround cost, over your horizon, outweigh the cost of building and running it.
Why automations fail in production: five causes
Automations can fail silently. Five causes (silent stops, duplicate writes, pasted secrets, runaway bills, app-level rules) and the fix for each.
How to use AI to extract invoice data safely
AI can read supplier invoices well enough to use. Human confirmation, validation and provider failover make it safe for a ledger. From a real trading ERP.
How much does business process automation cost?
Automation cost is driven by exceptions, systems that must agree, and rules that need a database. The formula, and what to measure for each term.
monday.com vs HubSpot CRM: pipeline or delivery
HubSpot CRM suits winning work (pipeline, attribution); monday.com suits delivering it. How to choose, how the bills differ, and when to run both.
Should you self-host n8n? What it really costs
Self-hosting n8n can cut a hosted bill but adds upgrades, an encryption key, backups and on-call. How to decide, with the break-even formula.
Zapier alternatives: who should leave, who should stay, and where to go
Most people seeking a Zapier alternative should move a few automations, not the account. How to decide by cost model and task shape, and where to go.
Make.com alternatives, and the cases where Make is still the right answer
Five good reasons to move off Make.com, where each one leads, and why most scenarios should be restructured before they are replaced.
Agency vs freelancer vs in-house for automation: who should build it
Agency, freelancer or in-house hire for automation: decide by cost of failure and rate of change, then test any builder with the same questions.
AI agents for operations: do you need an agent, or just a workflow?
Most operations problems need a workflow with a model at one step, not an agent. A decision test, plus the review, cost and audit controls agents need.
Airtable automation limits, and when to move to Make, n8n or code
Where Airtable's built-in automations stop, which symptom points to which limit, and when to hand the work to Make, n8n or a database with code over it.
Vibe coding to production: what has to change before real customers use the app
A checklist for taking an AI-built app live: access control, secrets, database rules, backups, tests, rollback, spend caps, and when to harden or rewrite.
Claude Code for business teams: setup, guardrails and what it is good for
How to set up Claude Code for a team: CLAUDE.md, permissions, hooks, MCP, CI and branch protection, what it suits in an ops codebase, and what to avoid.
Custom software, no-code or AI-built: how to choose
Compare SaaS, no-code, AI-generated code, custom engineering and a reviewed hybrid by process, data, permissions, lifespan, lock-in and exit path.
How to hire a developer who uses AI coding tools: questions to ask
Ten questions for a freelancer or studio using Claude Code, Cursor or Copilot, what good answers sound like, what to ask to see, and how to run a pilot.
What is an MCP server, and when does a business need one?
An MCP server lets AI tools use your systems through one controlled interface. What it exposes, when an API or Make scenario is enough, how to start.
Securing AI agents that act on business systems: prompt injection, permissions and review
Letting an AI agent read email and documents and act: prompt injection, least privilege, human approval, spend limits, logs and hostile-input testing.
How much does custom software cost? What drives the price
What sets the price of custom software: scope, roles, integrations, data and upkeep. What a quote covers, fixed vs time-and-materials, red flags.
How much does AI integration cost? What you pay for besides the build
AI integration cost is the build plus model usage, review labour, evaluation and upkeep. A cost-layer table, how to cap spend and how to size a pilot.
Claude API or OpenAI API for business automation: how to choose
Claude API or OpenAI API for business automation: test on your own data, check data terms, plan for model retirement and keep a fallback provider.
RAG, fine-tuning or long context: how to make AI answer from your documents
How to make AI answer from your documents: when to paste them in, when retrieval fits, when fine-tuning does not, and how to test and secure it.