# Internal tool or MVP: which one to build first in Dubai
An internal tool is software only your own team logs into: an admin panel, a booking board, a job tracker, a screen that pulls numbers out of three systems so nobody has to open three systems. An MVP is the smallest version of the thing your customers pay for. For most founders in Dubai the internal tool goes first, because it costs less, ships in weeks and produces evidence about your own business inside a month. We build internal tools and prototypes from AED 9,000 in about two weeks, and MVPs from AED 18,000, both at fixed scope and fixed price.
Three situations flip that default. They are worth reading before you spend anything.
I run slgo.ai out of Dubai. We build these things and then operate them, including the ones our own studio group depends on, so the numbers below are ours rather than borrowed from someone's survey.
What counts as an internal tool, and what counts as an MVP
The dividing line is who logs in and who pays.
An internal tool has a tiny, known user base. You already know their names. Nobody has to be persuaded to use it, because it replaces something they hate doing. Visual polish matters less than speed. Nobody churns. If it breaks on a Tuesday morning, you hear about it on Tuesday morning.
An MVP has strangers on the other side. It has to be findable, understandable without a walkthrough, safe behind a payment gateway, and stable when a user does something nobody predicted. Every one of those requirements costs real hours.
| | Internal tool / admin panel | MVP | |---|---|---| | Users | 2 to 20, all known to you | Strangers, unknown volume | | Question it answers | Where is my time and money leaking? | Will people pay for this? | | Onboarding | A ten-minute walkthrough | Must explain itself | | Cost of a bug | An annoyed colleague | A lost customer and a refund | | Payments | Usually none | Payment gateway, receipts, refunds | | Our price from | AED 9,000 | AED 18,000 | | Typical delivery | Around 2 weeks | 3 to 5 weeks by scope |
Same technology on both sides of that table. The AED 9,000 gap is the payment path, the defensive work around unpredictable users, and the simple fact that an MVP is judged by people who owe you nothing.
When the internal tool goes first
You already have revenue and something is leaking on the way to it. Enquiries land on WhatsApp, get copied into a spreadsheet, then into an invoice, and somewhere in that chain a follow-up gets forgotten. In a small Dubai service business at ordinary lead volumes, roughly AED 2,000 a month of salary goes to work a machine should be doing, and slow or unanswered replies cost around AED 4,100 a month in bookings that went to whoever replied first. Those are our own measured numbers. At that rate a build pays itself back inside four to six months, which is a shorter payback than almost any marketing spend.
Nobody in the company can state this week's paid revenue without a manual export. That one is a bigger problem than it looks. Founders who cannot see their own operating numbers make product decisions from memory, and memory over-weights the last angry client. An internal tool that shows paid bookings, source and response time on one screen changes what you argue about in Monday meetings.
Your product idea is downstream of your own operations. A marketplace built by someone who runs three of the businesses it will serve is a very different proposition from one built off a hunch. The internal tool becomes the data source that tells you what the product should do. Build it first and the MVP spec writes itself from real records instead of assumptions.
You are solo, non-technical, and this is your first time paying for software. The internal tool is a smaller bet on the relationship as much as on the product. You learn what a scope document feels like, what handover actually includes, and how it feels to have somebody else's code sitting inside your business, with AED 9,000 at risk instead of AED 60,000. I would rather a first-time client discover our working style on an internal build than on the thing their funding depends on.
Your team is the constraint and hiring is the alternative. An operations coordinator in Dubai costs somewhere around AED 5,000 to 8,000 a month plus visa and end-of-service. If a two-week build removes most of what that role does, the comparison is arithmetic, not philosophy.
When to skip straight to the MVP
Pre-revenue, idea-stage, no operations yet. An internal tool for a business that does not exist is a dashboard of zeros. Nothing to instrument means nothing to learn, and you will have spent AED 9,000 to confirm that. Go find out whether anyone pays.
A deadline you do not control. An investor conversation with a date on it, an accelerator cohort, a signed letter of intent from a first customer, a competitor about to launch the same thing. Time-to-market beats internal tidiness every time there is an external clock running. Ship the customer-facing thing, run your operations on WhatsApp and a spreadsheet, and accept the mess for a quarter.
The whole question is product-market fit and you have no evidence either way. Internal tools tell you nothing about the market. They tell you about your own process. If the open question is whether strangers will pay, only strangers can answer it, and only through something they can actually buy.
One caveat on that middle case. Plenty of founders declare a deadline that nobody outside their own head set. If the deadline is self-imposed, it is a preference, not a constraint, and it should not override the other signals.
Money and timelines in Dubai: the honest bands
Local development rates in the UAE sit at roughly AED 400 to 600 an hour for a Dubai-based team and around AED 90 to 180 an hour for offshore delivery with a local account manager. Those are reported market bands rather than anyone's published rate card, but the arithmetic is useful: an AED 40,000 quote at AED 450 an hour is under 90 hours of senior time, which is two working weeks for one person.
| Route | Internal tool | MVP | Timeline | What you carry | |---|---|---|---|---| | Freelancer | AED 3,000 to 8,000 | AED 8,000 to 25,000 | 2 to 8 weeks | Code that breaks once they stop answering | | Low-code / no-code, built by you | AED 70 to 900 a month | AED 200 to 1,500 a month | Your evenings for 4 to 10 weeks | A hard ceiling once the logic gets specific | | Fixed-scope AI-assisted build (ours) | From AED 9,000 | From AED 18,000 | About 2 weeks / 3 to 5 weeks | A narrow product, deliberately | | Development agency | AED 25,000 to 60,000 | AED 40,000 to 120,000+ | 3 to 6 months | Support at 15 to 25% of build cost per year |
Prices are approximate and worth confirming with each provider before you budget.
The agency row is not padding. A discovery phase, an architecture that survives ten times your current data, a QA pass, documentation and a security review are all real line items, and an agency delivering all of them for AED 90,000 is fair value. The question is whether your project needs those lines yet. A booking tracker for four staff does not.
Our fixed-scope route is narrower on purpose. Every screen, every action, every integration goes on a list before anything is written, and that list is the contract. What makes it something other than cheap contract work is the review gate: every line an AI model produces gets read by a human engineer before it touches customer data or a payment. Rapid prototyping without that gate is how you buy technical debt at speed. You can see the full terms and what sits inside a two-week fixed-scope build rather than reconstructing them from this table.
The internal tool is often automation wearing a different name
Roughly half the internal tool briefs that reach me do not need new software at all. They need the systems the business already pays for to talk to each other.
The pattern is consistent. There is a CRM. There is WhatsApp. There is an ad account. There is a payment gateway. Between each pair sits a human being copying a name, a number or a link from one window into another. Nobody designed that job. It accumulated.
Connecting those systems is workflow automation, and it is a different product with a different price: from AED 6,000 to set up one process and from AED 1,200 a month to keep it running, against AED 9,000 for a custom tool with a monthly hosting bill on top. When the process crosses tools you already own, wiring one process onto autopilot is the cheaper and faster answer, and it fails less often, because there is less new code in the world.
The WhatsApp side of this has a cost detail that trips people up. The free WhatsApp Business app cannot be connected to anything: no CRM sync, no automated payment link, no reliable multi-agent access. The WhatsApp Business API can, and it is billed per message. Meta's UAE marketing template rate is about USD 0.0499 per message, close to AED 0.18, and replies inside an open 24-hour customer service window have been free since late 2024. So a business sending 2,000 marketing templates a month pays somewhere near AED 370 in Meta fees plus whatever the provider adds. That is the real line between the free app and the connected one, and it is small enough that cost is rarely the reason to stay manual.
We run this on ourselves before selling it. The same contour handles the SkyLight studio group: an agent answers in under a minute at 3am on a Saturday and at 9am on a Monday with the same script, quotes, sends a payment link, and writes the deal into the CRM with its source tag attached. That operation clears about AED 67,600 net a month at a return on ad spend near 5x. There is no custom internal tool behind it. It is connected systems and a WhatsApp agent that qualifies and closes in the thread.
Three questions that decide it
| Question | Your answer | What to build first | |---|---|---| | Do you have paying customers today? | Yes, and operations hurt | Internal tool | | | No, still an idea | MVP | | Is there a deadline someone else set? | Yes: investor, cohort, signed LOI | MVP | | | No, the date is mine | Internal tool | | Does the painful work cross tools you already pay for? | Yes, it is copy-paste between systems | Neither. Automate first | | | No, it needs a screen nobody sells | Build |
Read the third row before the other two. If the answer there is copy-paste between existing systems, stop, because an AED 6,000 automation removes the pain and you keep the AED 9,000. When rows one and two disagree, the external deadline wins. A real date beats a good habit.
Can the internal tool become the MVP later?
Sometimes, and only if it was decided on day one.
The difference is structural. An internal tool assumes one company's data in one place, staff who are trusted by default, and a login list you maintain by hand. A product needs each customer's data isolated from every other customer's, self-service sign-up, an audit trail, and a billing relationship. Retrofitting isolation into a single-tenant data model is close to a rewrite, and it is the most expensive category of technical debt I see in this market.
Saying it out loud during scoping costs almost nothing. "This is internal now, and I want it to serve outside customers in eight months" changes the data model, the authentication approach and the deployment, all before a single screen exists. Said in month nine, the same sentence buys you a second project.
If you genuinely do not know, build the internal tool as an internal tool. A clean small thing that gets replaced is a better outcome than a bloated thing that was built for a future that never arrived.
How founders end up with two half-finished products
The failure is always the same shape, and it is never a technical failure.
A build starts as an internal booking tracker. In week one somebody says clients could log in and see their own jobs. In week two the login needs a password reset, which needs email, which needs a template. In week three it needs to take payments, so it needs a payment gateway, refunds and receipts. The internal tool is now an MVP that nobody scoped, and the MVP is now an internal tool that nobody can use. Both stall.
Fixed scope is the mechanism that prevents it, not a sales term. The list is written and frozen before work starts. Anything discovered afterwards goes on a second list with its own price and its own start date. That feels rigid in week two and saves the project in week four.
The sequencing rule I hold clients to: one build at a time, finished and in daily use for two weeks before the next one starts. Two weeks of real use reorders the second build's priority list more accurately than any planning session, and it costs nothing. Our published bands for both build types sit on the pricing page if you want to sanity-check the numbers above.
