Internal tools and data
Internal tools development for teams outgrowing spreadsheets
For founders, COOs, and ops leads whose operations now run on spreadsheets, copied data, and manual steps. We build the admin panels, ops portals, dashboards, and reporting pipelines your team works in every day, with permissions and audit history built in.
- Price
- Scoped on a call
- Model
- Priced by milestone
- Timeline
- Agreed in the proposal, phased for larger systems
What it is
What is internal tools development?
Internal tools development is the design and build of the software a company uses to run itself: admin panels, ops portals, dashboards, CRMs, data pipelines, and automated reporting. Pixelact builds them for startups and growing companies that outgrew spreadsheets or disconnected tools, so work happens in one system with clear permissions, audit history, and data people can rely on.
You need this when
Our operations run on a dozen spreadsheets, and nobody is sure which copy is current.
Someone on the team spends part of every week copying numbers between tools to build the same report.
Support staff edit records straight in the production database because there is no proper screen for refunds, plan changes, or approvals.
We pay for a CRM that fits half our process, and the other half lives in side spreadsheets and Slack threads.
We cannot tell who changed a record or approved a payment, and investors will ask during due diligence.
What you get
What you get at the end
Every deliverable is written down in the proposal before work starts, so you know exactly what you are buying.
- 01
A working tool in production
The admin panel, ops portal, dashboard, or CRM agreed in the proposal, deployed to your own cloud accounts and released to the team it was built for.
- 02
Roles, permissions, and audit history
Access rules for each role, enforced on the server, and a searchable log of who changed what, when, and from which value.
- 03
Data model and migration
A documented schema for the records the tool owns, and a tested import of your spreadsheet or legacy data, with every rejected row listed and the reason it failed.
- 04
Integrations and pipelines
Connections to the systems you keep, such as billing, your CRM, or your product database, with scheduled jobs, retries, and an alert when a sync fails.
- 05
Automated reports
Reports and dashboards that build themselves from source data on a schedule, with every metric defined in writing so finance and ops read the same number.
- 06
Handover pack
The repository, runbooks, architecture decision records, and a walkthrough with whoever takes over, so any engineer can maintain and extend the tool.
How it works
How it works, step by step
A written process with a named lead, so you always know what happens next and who owns it.
- Step 01
Discovery call
A 45-minute call about how the work runs today, what it costs you, and what a good outcome looks like. No slides, and no solution until the problem is clear.
- Step 02
Proposal walkthrough
Within 48 hours we present scope, exclusions, milestones, two or three options, and payment plans on a live call, then confirm everything in writing.
- Step 03
Workflow and data mapping
We sit with the people who do the work, map each step and record, and agree roles, permissions, and metric definitions before designing screens.
- Step 04
Build in milestones
We ship working slices to a preview environment, run real data through them early, and review each milestone with the people who will use the tool.
- Step 05
Migration and rollout
We import existing data through the tool's validation rules, move all edits to the tool on an agreed date, and reconcile totals against the old reports through one full business cycle.
- Step 06
Handover
Code and final files transfer once the final invoice is paid, with runbooks and a walkthrough for whoever takes over. The accounts were yours from the start, and a retainer is offered once the tool has proved its value.
What does internal tools development cover?
#Internal tools development covers the software your own team uses to run the business, as opposed to the product your customers buy. Most projects combine two or three of the pieces below around one workflow, such as onboarding a customer, approving a refund, or closing the month.
- Admin panel
- A back office for your product: look up an account, change a plan, issue a refund, or see what a customer sees, with each action limited by role and logged. Custom admin panel development makes sense once support or ops staff need production access to do their jobs.
- Ops portal
- One place where an operations team works a queue: intake, assignment, approvals, status, and SLA timers, replacing the shared inbox and the tracking spreadsheet.
- Dashboard
- A screen that answers one recurring question for one role, such as which accounts are at risk this week. Custom dashboard development starts from the decision the screen supports, then picks the charts.
- Custom CRM
- A customer or pipeline system shaped around how you sell or deliver. Custom CRM development is worth it only when a mainstream CRM forces more workarounds than it removes.
- Data pipelines and automated reporting
- Scheduled jobs that pull data from billing, product, and sales systems into one database, then produce the weekly, board, or investor reports nobody should assemble by hand.
Should you build or buy an internal tool?
#Buy when the process is standard and a product already fits most of it. Build when the workflow is specific to how you operate, the data spans several systems, or the access and audit rules go beyond what a product or low-code builder can enforce. When buying is the better answer, we say so on the discovery call.
| Option | Good fit when | Trade-off |
|---|---|---|
| Off-the-shelf SaaS, such as a mainstream CRM or helpdesk | Your process matches how the product expects you to work | You adapt to its data model, and side spreadsheets grow around the gaps |
| Low-code builder, such as Retool or Airtable | A small team needs working screens over data that already exists | Per-seat pricing, platform limits, and logic spread across many small apps |
| Hire an in-house engineer | Internal systems will need full-time work for years | Hiring is slow, and one person can become the only one who understands the system |
| Custom build with a studio | The workflow is core to operations and needs real permissions, audit history, and integrations | Higher upfront cost, and someone has to own maintenance after handover |
A common path is a spreadsheet or low-code tool first, then a custom build once the workflow stops changing every week and the workarounds cost more than the build. Our guide to replacing spreadsheets with internal tools walks through the signals.
How do permissions and audit history work in a custom admin panel?
#Permissions decide who can see and change each record, and audit history records every change after it happens. We design both before the first screen, because adding them to a tool people already use means reworking every query and form.
Permissions enforced on the server
#We define roles from the real jobs on your team, such as support agent, ops manager, and finance approver, and deny everything by default. Checks run in server code, not only in the interface, because a hidden button does not stop a crafted request. Broken access control is ranked first in the OWASP Top 10:2025.
Where records are split by region, team, or client, we add Postgres row-level security. The application passes the signed-in user to each transaction, and the database returns only the rows that user may see, even if an application check is missed. The application connects as a role that does not own the tables, because superusers always skip these policies and table owners skip them unless the table forces them.
Audit history people can use
#Every change writes an entry with who acted, what changed, the old and new values, and a reason when the action needs one, such as a refund or a manual balance adjustment. The entry is written in the same database transaction as the change, so neither can exist without the other. The application can add and read entries but never edit or delete them.
create table audit_log (
id bigint generated always as identity primary key,
occurred_at timestamptz not null default now(),
actor_id uuid not null,
action text not null,
entity text not null,
entity_id text not null,
previous_value jsonb,
new_value jsonb,
reason text
);
create index audit_log_entity_idx
on audit_log (entity, entity_id, occurred_at desc);
grant select, insert on audit_log to app_user;Ops leads get a searchable history on each record, and finance gets an export for reviews. The startup security checklist lists the access and logging checks to run before customers or investors ask.
How do you keep data quality high when moving off spreadsheets?
#Data quality comes from validating every record at entry and giving each field one place where it is edited. A new tool fed the same messy inputs produces the same wrong report faster, so we plan the migration as carefully as the screens.
- 01Inventory the sourcesList every spreadsheet, export, and tool that holds part of the workflow, who edits each one, and which columns are calculated rather than typed.
- 02Pick one system of record per fieldDecide where each value is created and edited, such as plan and price in billing and account owner in the new tool. Every other place reads it from there.
- 03Enforce the rules in the databaseRequired fields, allowed values, and rules that span fields go into Postgres constraints as well as the form, so no import, script, or bug can skip them.
- 04Import through the same rulesExisting rows pass the same checks. Each rejected row lands in a report with its reason, and the team fixes the data at the source before the next run.
- 05Switch edits, then reconcileFrom an agreed date, records change only in the tool, and the old sheet becomes a read-only export. Through one full cycle, such as a month-end close, compare totals with the old report until they match.
What do data pipelines and automated reporting involve?
#Data pipelines copy records on a schedule from the systems you already run, such as Stripe for billing, your sales CRM, and your product's own database, into one reporting database. Automated reporting builds the weekly, board, or investor report from that copy, so nobody assembles numbers by hand.
We write each metric definition down before building the chart, so active customer or monthly recurring revenue means the same thing in every report. Each job is safe to rerun, retries on failure, and alerts a named person when a sync stops.
We usually start with Postgres and scheduled jobs, reading product data from a replica so heavy report queries do not compete with customer traffic. A warehouse comes in only when data volume or query load calls for one.
What makes a custom dashboard useful?
#A custom dashboard is useful when people trust its numbers enough to act on them. Each one we build shows when its data last synced and lets a reader open the records behind any figure that looks wrong. Row-level permissions apply here too, so a regional manager sees one region while finance sees the whole company.
How is internal tools development priced?
#Internal tools builds are priced by milestone after a discovery call, never from a rate card. The proposal lists scope, exclusions, milestones, and two or three options. Larger systems are split into phases, each with its own milestones.
- Start small. A pilot sprint starts from $4,500 and can automate one workflow or prove one integration before a larger commitment.
- Review an existing tool first. A software audit starts from $2,500 and ranks the risks and fixes. If it leads to a build within 60 days, half the audit fee is credited against the build.
- Choose a payment plan. Each quote offers two or three plans, such as 50% at start and 50% on delivery, or 40%, 30%, and 30% across start, working midpoint, and launch.
- Pay providers directly. Hosting, licenses, and paid APIs sit in your accounts, and any cost we front is passed through at cost.
How should you prepare for the first call?
#You do not need a specification. Bring enough to show how the work runs today.
- The spreadsheets, tools, and inboxes the workflow touches, including the unofficial ones.
- One recent example followed from start to finish, including where it stalled.
- Who uses it, and what each role should and should not be able to do.
- The reports people depend on, who reads them, and how long they take to build.
- Any approval, retention, or review rules your finance team or customers expect.
- The person who will own feedback and sign off each milestone.
If you can, bring whoever makes the final call, and a link to any tool or export the team already uses. Our project process explains the stages after the call, and our engineering standards cover review, testing, and handover.
FAQ
Common questions
Straight answers on scope, price, ownership, and what happens next.
It depends on how many workflows, roles, and integrations the tool covers, and on the state of the existing data. We give a timeline with milestones in the proposal, after the discovery call, rather than a number up front. If the scope is uncertain, a pilot sprint can automate one workflow first and turn the larger estimate into something you can check.
Yes. Most internal tools read from SaaS products you already pay for through their APIs, and from your own product database through a read replica. We avoid writing straight into another product's tables, even where access allows it, because its vendor can change the schema without notice. Where a tool must write back, it goes through the official API with retries, idempotency keys where the API supports them, and a logged result.
Sometimes, yes, and the choice is less final than it looks. If the data lives in a Postgres database you own, a low-code front end can serve a small team now and be replaced by custom screens later without migrating records. Some teams keep low-code for read-only views and move the audited write paths, such as approvals and refunds, into custom code first.
You do. The repository, cloud accounts, and databases sit in your name from the start, and the code and IP transfer to you when the final invoice is paid. Handover includes runbooks, decision records, and a walkthrough with whoever takes over, so another engineer, in-house or external, can pick the tool up without us.
Yes, where it saves real time and a person can check the result. Common examples are sorting incoming requests, pulling fields out of documents, and drafting replies for review. AI features start with a pilot that measures accuracy on your own data, and the proposal states data boundaries and what happens when the model is wrong. See AI systems for that work.
Internal tools change as the business changes, so most teams want someone on call. After a build proves its value, we offer a retainer with fixed monthly capacity. Maintain gives 20 hours a month with a 2 business day response, and Improve gives 50 hours a month with a 1 business day response. The minimum term is 3 months. See how the engineering retainer tiers compare.
No tool makes a company compliant on its own. Compliance depends on your policies, people, and processes as well as your software. We build the parts software can carry, such as role-based access, audit history, data export, and retention rules, and we document them for your auditor or advisor, who confirms whether your controls meet the standard.
Yes. Most people who use internal tools are not engineers. A named lead runs the project, maps the workflow with the people who do the work, and reviews each milestone with them in plain language. Decisions are written down, so the ops lead and the founder see the same scope, trade-offs, and open questions.
Related
Where to go from here
Related services, guides, and checklists.
Start a project
Tell us what you're building.
Share a few details in a short form, or book a discovery call if you're not sure where to start.
On the discovery call
- 01What prompted this now, and where things stand
- 02What a good outcome looks like
- 03Whether we're a fit, and what a next step could be
45 minutes. No slides, and no long questionnaire.