Automation

Internal Tool & Dashboard Development

When the work is running on six spreadsheets, a WhatsApp group and one person's memory, the fix is usually a small internal tool: one place to see the jobs, who owns them, what is late and what is unbilled. I build those on top of the data you already have, sized to a team rather than to an enterprise.

  • Internal ops and task tools built for an agency in Kathmandu
  • Airtable, Notion, Sheets, or custom where needed
  • Your data, your accounts

Who this is for

Teams of roughly five to fifty people whose operations have outgrown spreadsheets but do not fit the software built for their category, or whose staff quietly stopped using the expensive system because it demanded twenty fields to record one job. Agencies, clinics, consultancies, manpower firms, construction and trading companies all reach this point. Internal tools are one of the builds under marketing and business automation services.

It is not for a team of three who need a shared list. That is a spreadsheet with two rules, and I will say so. Nor is it for replacing accounting or payroll software, where the compliance requirements make an off-the-shelf product the right answer.

What I build

  • Task and job systems. Work items with an owner, a due date, a stage and a client, visible as a board or a list, with automatic reminders on what is overdue and a weekly digest of what each person has coming.
  • Client and project dashboards. One screen per client: active work, last contact, spend, deliverables due, outstanding invoices and the documents that matter, assembled from the CRM, sheets and ad accounts rather than typed in.
  • Operations trackers. Inventory, deliveries, applications, cases, admissions or candidates, with the stage transitions your process actually has and the notifications each stage should trigger.
  • Approval systems. Requests for leave, purchases, discounts or content, routed to the right approver with a record of who approved what and when, which is usually what an audit is missing.
  • Reporting screens that combine sources nobody wants to open separately, built as described on automated marketing reports and dashboards.
  • Internal AI helpers where useful: a search over your own SOPs and past work, covered on custom GPT and AI assistant development.

What I deliberately do not build

  • Accounting, payroll and statutory filing systems. Use software built for the local compliance requirements and connect to it instead.
  • Customer-facing applications. This service is internal tooling. Public products need a different level of security review, testing and support than a team tool, and that is web development rather than automation.
  • A replacement for a system people already use happily. If the CRM works, the tool wraps around it.
  • Anything that stores card numbers or identity documents without a proper security design and your explicit decision.
  • Feature lists collected from a workshop. The first version does the three things that hurt today. Everything else waits for evidence it is needed.

How a build runs

  1. Watch the work for a day. Not a requirements meeting: the actual sequence, the spreadsheets people keep privately, and the moments where someone asks a colleague a question the system should answer.
  2. Choose the platform honestly. Airtable, Notion or Google Sheets with Apps Script cover most cases at low cost and let your team change things themselves. A custom application is only worth it when the data model, the user count or the permissions genuinely exceed those tools, and I will tell you which side you are on.
  3. Model the data first. Entities, relationships and the one thing each record is the truth about. Skipping this is why the second version of most internal tools is a rebuild.
  4. Build a usable first version in two to three weeks and put it in front of the people who will use it, with their real data migrated.
  5. Iterate on observed use. What they avoid, what they still keep in a private sheet, and what they ask for once it exists. Two short rounds, not an open-ended queue.
  6. Hand over with documentation, permissions set, a backup schedule and training, plus the ability for a named person on your side to make simple changes without me.

What you receive

  • The tool in your own accounts and workspaces, with the data in your own storage and exportable at any time.
  • The data model diagram and a written description of every automation attached to it.
  • Role-based permissions, so people see what they need and not the salary column.
  • Backups, an error alert channel and a documented recovery step.
  • Documentation, a recorded training session and a short guide for adding a field or a view without breaking anything.
  • Optional monthly support for changes as the process evolves.

Proof

I have built internal operations and task-management tools for an agency in Kathmandu and for my own work: job tracking, client dashboards, approval flows and reporting screens on top of Sheets, Airtable and automation platforms. I can walk through them on a call rather than describe them here, since they hold client data.

Pricing pointer

Setup is quoted on request; a single tracker with automations sits at the lower end of the range. A multi-entity operations system with dashboards, approvals and integrations to your CRM and ad accounts sits at the top. Monthly support is quoted on request; platform subscriptions such as Airtable or Notion are paid in your own name. Where the need is really connecting existing tools rather than a new screen, that is workflow automation, and sheet-based data work sits under Google Sheets, document and data automation.

Next step

Send a screenshot of the spreadsheet your operations currently run on and tell me the question you most often cannot answer quickly. I will come back within four business hours with whether this needs a tool or a tidy-up, the platform I would use and a fixed price.

Frequently asked questions

Should we build a custom tool or buy software?

Buy, whenever a product fits the process without heavy customisation, because you get support and updates. Build when the process is genuinely specific to how you make money, when the per-seat cost of a product is out of proportion to the few features you need, or when the product exists but your team refuses to use it. I say which applies before quoting.

What does it run on?

Usually Airtable, Notion or Google Sheets with Apps Script, connected through n8n, Make or Zapier. These are cheap, fast to change and can be maintained by someone on your team. A custom database and application is used only when user count, data volume or permission complexity make the low-code route a false economy.

Who owns the tool and the data?

You do. Everything is built in accounts registered to your company, the data stays in your storage and can be exported in a standard format at any time, and the documentation is written so another developer can take it over. There is no hosting fee to me and no lock-in.

How long does it take?

A usable first version in two to three weeks for a single tracker, four to eight weeks for a multi-part operations system including data migration. The pace is usually set by how clean the existing data is and how quickly your team can review the first version, not by the building.

What happens when we need changes later?

Most changes such as adding a field, a view or a filter are designed to be done by a named person on your side, and the handover training covers exactly that. Structural changes, new integrations or new automations are covered by monthly support or quoted individually. Nothing about the tool requires me to keep working on it.

Can it connect to our CRM and ad accounts?

Yes. That is usually the point: the tool becomes the one screen where CRM records, ad spend, sheet data and job status appear together. Connections are read-only where possible, and where the tool writes back to the CRM, the rules about which system owns which field are agreed before the build.

Ready to talk about your project?

A free 30-minute call, a straight answer about what would move the numbers, and a written proposal within 48 hours if we are a fit.