What you're buying: a defined workflow that runs in your n8n instance or agreed hosting environment, uses your provider accounts, handles failures visibly and comes with enough documentation for another developer to operate it. For role selection, costs and what should be automated, see the AI automation expert page. For connected marketing and CRM workflow design, see AI marketing automation for business.

AI automation is one part of my broader practice as a digital marketing expert in Nepal — it sits alongside SEO and paid media work, most often automating the lead handling and reporting those channels generate.

Scope and deliverables

Before build, we agree the trigger, systems, fields, rules, exceptions, approval points, expected volume and acceptance checks. A typical delivery includes:

  • Workflow map and agreed boundary. What enters, what leaves and what remains manual.
  • Production n8n workflow. Integrations and custom code needed for the agreed process.
  • Reliability controls. Validation, duplicate protection, retries, timeouts, fallback paths and failure alerts.
  • Test and acceptance record. Normal cases, malformed input, provider failure and human approval paths.
  • Operations documentation. Credentials used, configuration, schedules, recovery steps and known limits.
  • Handover. Workflow exports and a walkthrough for the owner or technical contact.

Anything outside that boundary is a change in scope, not an invisible extra. Provider subscriptions, hosting and API usage are paid directly through your accounts unless we explicitly agree otherwise.

What I build

Lead capture and routing

Enquiries arrive from a website form, a Facebook lead ad, WhatsApp and email, and then sit in four different places until somebody remembers to check them. The automation consolidates them into one pipeline, deduplicates, enriches where useful, classifies the enquiry, routes it to the right person and triggers a first response inside minutes rather than hours. Response time is usually the single most valuable thing to fix in a small sales operation.

Document and data extraction

Invoices, purchase orders, CVs, bank statements and delivery notes arrive in inconsistent formats. A model is genuinely good at pulling structured fields out of that mess. The important engineering isn't the prompt, it's the validation: schema-constrained output, sanity checks on the extracted values, and a human review queue for anything below a confidence threshold.

Reporting pipelines

Pulling numbers from ad platforms, analytics and a CRM into one scheduled report, with the anomalies flagged in plain language rather than buried in a dashboard nobody opens. This is mostly deterministic work with a thin summarisation layer on top.

Support triage and drafted replies

Incoming tickets get classified, tagged, prioritised and matched against your existing documentation, and a draft reply is prepared for a human to approve or discard. I deliberately default to draft-and-approve rather than full auto-send, because the reputational cost of a confidently wrong automated reply is much higher than the time saved by skipping review.

Internal tools and integrations

Where an off-the-shelf integration doesn't exist, I write it. That is the practical advantage of n8n over the no-code platforms: when you hit the edge of what the visual nodes support, you drop into code inside the same workflow rather than abandoning the platform.

How I build it so it doesn't break

An automation that works in a demo and fails silently in production is worse than no automation, because people stop checking. Every workflow I deliver includes:

  • Structured output validation. Model responses are constrained to a schema and validated before anything downstream acts on them.
  • Confidence thresholds and human approval. Anything that touches money, customers or public communication gets a review step.
  • Retries, timeouts and fallbacks. Third-party APIs fail. The workflow has to expect it.
  • Run logging and failure alerts. If a workflow stops working, somebody finds out that day, not next quarter.
  • Cost control. Token usage is estimated per run before build and monitored after, so the automation stays cheaper than the work it replaces.
  • Idempotency. Re-running a workflow must not double-charge, double-send or double-create.

The broader method, including how I decide what to automate first, is written up in this practical automation playbook, and there's a worked example of an outbound pipeline in the lead generation system breakdown.

The stack

n8n for orchestration, self-hosted or cloud depending on your data sensitivity and volume. OpenAI API for language tasks, with the model chosen per task rather than defaulting to the largest one. Python or TypeScript for anything the nodes can't express. Vector storage where retrieval over your own documents is genuinely required, which is less often than vendors suggest. Everything version-controlled and exported so it's portable.

How engagements work

The first step is a short discovery conversation to find the processes that are repetitive, high-volume and rule-shaped enough to be worth automating. Not everything is. I'd rather tell you a process is a bad automation candidate than take money to build something fragile.

From there it's usually one narrow workflow built end to end first, so you can see real value before committing to more. Broad "automate the whole business" projects fail; narrow ones that save a specific team a specific number of hours per week succeed and then expand naturally.

Ownership: you keep the n8n instance, workflow exports, API keys, custom code and documentation. Support: an agreed defect in the delivered scope is different from a provider changing its API or a new requirement. Ongoing monitoring, upgrades and changes can be handled under optional support or by your team; neither is a condition of handover.

What it costs

My prices are published: an audit with a 30 day action plan is USD 249, a four week growth sprint is USD 899, and ongoing work is USD 1,800 a month, which is roughly NPR 34,000, NPR 122,000 and NPR 245,000 a month at current rates. The exact NPR figures and what each tier leaves out are on the pricing page, and the first thirty minute call is free. Published results, with baselines and a note on which were agency team engagements, are on the case studies page.

Common questions

What is n8n and why use it instead of Zapier or Make?

Self-hosting, arbitrary code inside workflows, and no per-task pricing at volume. If your needs are simple and low-volume, Zapier is often the better answer and I'll tell you that rather than over-engineer it.

What happens when the AI gets something wrong?

It will, sometimes. That is why consequential steps are gated behind human approval and every output is validated against a schema before anything acts on it.

Is my data sent to OpenAI?

Only what a given step needs, and only if a language model is actually involved in that step. Where data sensitivity rules that out, the workflow is designed to keep processing local, or the model is dropped from that step entirely.

Do you provide ongoing maintenance?

Optional. APIs change and workflows drift, so a small maintenance arrangement is usually sensible, but it isn't a condition of the build and you're free to maintain it in-house.

What is included in handover?

The workflow in your instance, exports, relevant custom code, configuration and recovery notes, acceptance checks and a walkthrough. Your API keys remain in your accounts.

Start with the expert, specialist, consultant and cost guide if you're deciding what to automate. See connected AI marketing workflows for attribution, CRM scoring, nurture, audience sync and consent. If the project needs a custom application around the workflow, see web development.

Send one workflow for a scoped answer

Name the trigger, tools, monthly volume, desired result and the exceptions your team handles manually. I'll respond with the likely boundary, open questions and whether n8n is the right fit.

Get in touch Review costs and fit