# Expense management for Dubai SMBs: cards, apps and where the approval chain breaks
TL;DR: For a Dubai company of 5 to 30 people, expense management breaks down between the card swipe and the ledger, not inside any single app. Published prices run from free (Pemo Starter, Pluto Starter for up to 10 card users, Mamo Growth) through AED 99 and AED 399 a month (Mamo Premium, Pemo Premium) to AED 1,500 at the enterprise end, while Alaan and Qashio quote per company. The cheapest fix is usually putting the approval where the approver already looks, not buying a new card.
The same failures show up in almost every Dubai SMB I look at, in the same order. A receipt gets photographed three weeks after the purchase, or never. The one person who can approve a spend is out of the country for two weeks in August. The same courier charge gets claimed twice, once on a card statement and once as a cash claim. A contractor who paid for materials on a personal card is still waiting forty days later, the kind of thing that quietly costs you the contractor.
None of that is a software gap. Card platforms in the UAE have solved the capture part well for years. The gap is the handoff between steps, and handoffs are what nobody buys software for. I am also not going to quote a percentage of UAE SMEs that mishandle expenses: those figures circulate between vendor blogs without a primary source, and a number you cannot trace is worse than no number.
Where the chain breaks in a 5 to 30 person company
Your flow has the same shape as everyone else's: someone spends, a receipt is captured, somebody approves, the transaction gets coded to a cost centre, it lands in the books. Every SMB I have mapped in Dubai has all of those steps and no owner for the gaps between them.
The receipt gap is the most visible. A salesperson pays for client lunch at 2pm, the paper receipt goes into a wallet, and the wallet empties on the 29th. By then the amount is legible and the reason is not, so the real cost is approving a spend nobody can evaluate any more.
The approver gap shows up seasonally. July and August empty the approval chain in Dubai, and Ramadan compresses working hours, so a queue that normally clears in a day takes three. If one named person is the only route to a yes, your spend process has the same availability as their phone.
The duplicate gap is quieter. One expense exists in two systems with two different shapes: a card line in the statement, a manual claim in a spreadsheet. Nobody is stealing anything, the controls simply cannot see both at once. Subscriptions do it slowly: one duplicated design tool at AED 150 a month is AED 3,600 over two years that never came up for approval again.
The reimbursement gap has a human cost. A junior or a contractor who fronts AED 2,400 for a shoot or a site visit is running an unsecured loan to your company. They rarely escalate, they just stop volunteering to pay for things.
What the broken chain actually costs
The admin hours are real, and smaller than vendors imply. A Dubai office coordinator on an AED 4,500 package costs about AED 5,441 a month once the visa cycle, insurance and gratuity accrual are counted, roughly AED 31 per working hour over a 22-day month of 8-hour days. Chasing 12 missing receipts at 8 minutes each is 96 minutes, month-end matching of unmatched card lines another 3 hours, so direct labour comes to under AED 150 a month. That alone is not a business case.
The costs that matter sit next to it. If card transactions reach the books in a monthly batch, every spend decision is made against data 15 to 45 days old, exactly the window where subscription creep goes unnoticed. The reimbursement lag has a price in goodwill that never appears on a P&L. And uncategorised spend cannot be cut, because you cannot reduce a tools budget you have never seen as a line.
Size it on your own data. Pull three months of transactions, then count how many reached the books more than 15 days after the purchase date, and how many carry no cost centre. Both counts above a third of your volume means a platform pays for itself. Both near zero, with a team that only complains about the app, means your problem is workflow design, not tooling.
Card platforms vs expense management apps, and what each one charges
"Expense management software" in the UAE is a label worn by several different kinds of company, and the confusion costs people money.
One group issues the card and runs the expense management software on top of it. Pemo, Alaan, Qashio, Pluto and Mamo sit here. You get physical and virtual cards, limits per card, receipts captured against transactions, and coded data pushed into your accounting tool. Card and workflow arrive as one product, so setup is fast and switching later hurts.
Worth knowing who stands behind those cards, because it is not always the brand on the front. Alaan's UAE cards are issued by NymCard Payment Services LLC, authorised by the Central Bank of the UAE under the Retail Payment Services and Card Schemes Regulation, with Visa as scheme partner. Pemo, founded in 2022 and stating more than 6,000 UAE businesses, received in-principle approval from the Central Bank for a Stored Value Facilities licence in July 2026. A vendor who cannot say in one sentence who holds the licence and where your float sits has answered your first question badly.
Another group never touches your money. Zoho Expense and Expensify reconcile statements from whatever card your bank already gave you, run the approval flow, and code the result. No new banking relationship and no float, but also no control at the card level, so limits get enforced afterwards by a human saying no.
Then there is your bank. Wio, Commercial Bank of Dubai, and traditional business accounts and corporate card programmes from names like American Express give you the card, the statement and, in newer digital accounts, virtual cards. Multi-step approval with category routing is what they leave to you.
Rate cards here move, so treat the following as what each vendor was publishing at the time of writing and check the live page before committing.
- Pemo publishes plans openly: Starter with no SaaS fee, Premium at AED 399 per company per month, Enterprise from AED 1,500 per month, with unlimited users and unlimited physical and virtual cards on each tier (Pemo plans).
- Pluto prices in dollars: Starter at USD 0 for up to 10 card users with unlimited virtual cards, Growth at USD 99 a month for up to 30 users plus USD 9 per extra user, Enterprise on quote (Pluto pricing).
- Mamo Business runs Growth free, Premium at AED 99 a month, Enterprise on quote above AED 500,000 processed monthly. Card spend carries 1.9% plus AED 0.80, collections above your card spend 2.7% plus AED 0.80 (Mamo pricing).
- Zoho Expense has a free tier for up to 3 users with 20 receipt autoscans, then Standard at USD 4 per user per month and Premium at USD 7, with card reconciliation and multi-level approvals from Standard up. Even the free tier connects to Zoho Books, QuickBooks Online and Xero (Zoho Expense pricing).
- Alaan publishes a tiered pricing page without a flat monthly figure per tier. It has run a 2026 offer of twelve months free on Alaan Premium, valued by Alaan at AED 6,000, plus up to 2% cashback on international spend on Premium and Enterprise (Alaan pricing). Those are the vendor's own numbers.
- Qashio prices per company with a free trial rather than a published rate card (Qashio pricing).
- Wio is the bank-side comparison point. Third-party guides put its Grow plan near AED 249 a month, but the document to read is Wio's own key fact statement rather than a blog.
For a 20-person company the decision turns on pricing shape, not headline price. Pemo Premium at AED 399 covers unlimited users. Pluto Growth at USD 99 is roughly AED 363 for up to 30. Zoho Expense Standard at USD 4 per user is about AED 15 a head, so 20 people lands near AED 295, without a card. At five users the per-seat model wins easily; at twenty-five it usually does not.
Approval rules that do not stall
Approval flows in Dubai SMBs fail because everything needs a human and only one human exists.
Set an auto-approve threshold from your own data, not a template. Sort three months of transactions by amount and find the value below which about 80% of your spend events sit. In the companies I look at that lands between AED 300 and AED 500 for general staff. Everything under it clears on the card with a receipt requirement and no human step, giving the remaining approvals a chance of being read properly.
Give every approver a named deputy and make the handover automatic: no decision in 48 hours, the request moves by itself. Escalation that depends on the requester chasing someone is a habit, not a rule, and habits die in August.
Route by category once you pass roughly 15 people. Operations spend, marketing tools and client entertainment behave differently enough that one approver cannot sanity-check them all, and marketing subscriptions in particular need someone who knows what the company already pays for.
Set limits at the card, not in policy documents. A per-transaction cap and a monthly cap per role means the control happens before the money moves, the whole argument for a card-issuing platform over a reconciliation tool.
Receipts, and what happens when there is no receipt
The capture rule that works is a photo at the point of purchase, before the card is back in the wallet. Every platform here supports it from a phone, and every platform fails when the habit is missing.
Optical character recognition handles the merchant, the amount, the date and the currency. What it cannot infer is why you bought the thing and which project it belongs to, the field your accountant needs. So capture keeps a human component: one line of purpose, one cost centre. Ask for two fields and people comply; ask for six and they will not.
Write the missing-receipt path before you need it, and make the declaration count per person visible rather than punitive (the FAQ has the version I use). What you are protecting is simple: a card line with an image attached, a stated purpose, a named approver and a timestamp is a complete record of a decision, and that is what lets you answer questions about your own spend in minutes.
Where the approval should actually live
Here is what the software category gets wrong at this size. The approval flow lives inside an app, and the app assumes the approver opens it. For a founder or an operations manager in Dubai that assumption is false. They live in WhatsApp, and an approval waiting for an app to be opened waits.
Pluto has clearly reached the same conclusion, since it accepts receipt submission over WhatsApp and syncs the image to the matching transaction. That is the right instinct applied to half the problem. The approval itself is the half that still stalls.
We build the other half, and the mechanics matter more than the idea. A transaction event fires from the card platform; the approver gets one message in the thread they already read, carrying merchant, amount, spender and category; a single-word reply writes the decision back into the platform. Send one transaction per message, never a digest: batch five items behind an "approve all" and you have built a button that approves things nobody read. Bind every reply to a specific transaction identifier, because a "yes" arriving four minutes after two notifications is ambiguous, and ambiguity in a money flow is a defect. Make the write-back fail loudly in the thread too, because a chat approval that never lands in the platform leaves a decision living only in a conversation.
The boundary worth naming: slgo does not issue cards, is not an accounting firm, and does not resell Pemo, Alaan, Zoho or anyone else. We wire the tools you already chose to each other so a spend event reaches a decision and a decision reaches your books without a person carrying data between screens.
Getting the card feed into your books without double entry
This is the step people leave for later and never do, which is how a modern card platform ends up producing a spreadsheet that gets typed into accounting software by hand.
Check the connector before you buy the card. Pemo's own material lists QuickBooks, Xero, Zoho Books, Wafeq and Tally among its integrations, and Zoho Expense connects to Zoho Books, QuickBooks Online and Xero from the free tier. If your accounting tool is a desktop licence on one machine, assume a connector layer and price it in.
Choose real-time sync or a monthly batch deliberately. Real-time means your books carry today's spend and your accountant reviews rather than retypes. Under 30 people it matters most on cards and least on recurring vendor invoices, which arrive on a predictable cycle anyway.
Code to the cost centre at capture. If the spender picks the project from a short list in the moment, the coding is right and free. If your accountant allocates 200 transactions retrospectively from memory, the coding is a guess with a category name attached, and your project profitability inherits the guess.
Keep contractors off the company card and give them a clean path instead: a document in, a payment link or transfer out on a fixed weekly run. What fixes reimbursement lag is a fixed payout day, not faster approval, because a fixed day requires nobody to be available. Underneath all of it sits one rule: one system of record per fact. The card platform owns the transaction, the accounting tool owns the ledger entry, the chat thread owns nothing. Every reconciliation nightmare I have been called into started with two systems both believing they owned the same fact.
A week is enough for the first version
- Days one and two, policy on one page: the auto-approve threshold, who approves what, the deputy for each approver, the cost centre list, the missing-receipt rule.
- Day three, cards. Virtual for subscriptions and online spend, one physical per person who travels, limits set per role at issue rather than later.
- Day four, the accounting connection, tested with a real AED 20 purchase carried from capture through approval into the ledger.
- Days five to seven, a pilot with the three people who spend most often, then rollout with the old process switched off rather than left running alongside. Parallel processes are how companies end up with the duplicate problem they bought the platform to solve.
What fails in week two is rarely the software. It is usually that nobody told the team the reimbursement day is Thursday, so they keep asking.
How we run this
Our own contour runs the same shape on the sales side of the group: an enquiry lands in the CRM as a deal with its source tag, a payment link is generated against that specific booking, and the paid status flips without anyone copying anything. That pipeline currently runs a studio business at AED 67,600 net a month at roughly 5x ROAS. Our expense management setup uses the same rails in reverse, with an approval notification going out instead of a payment link coming in.
Putting one process like this on autopilot starts at AED 6,000 to build and AED 1,200 a month to run and watch, connecting the tools you already pay for rather than selling you new ones: details on AI automation. If the same thread also handles customer conversations rather than only internal approvals, that is the WhatsApp AI agent, and everything we charge sits on our pricing page. To see the arithmetic on your own transactions first, a Growth Audit starts at AED 3,000 and returns the age of your spend data, the approval bottleneck by name, and the flow we would wire.
Start by exporting last month's card transactions and counting how many reached your books without a human retyping them. That number tells you whether you are buying a card, a workflow, or neither.
