There are four common ways to resource a software build, and most decisions about which one to use are made in the wrong order. The discussion starts with cost or familiarity, "we have always used agencies", "we have headcount budget", "offshore is cheaper", and a model gets attached to the work before anyone has characterised the work itself. The result is the recurring pattern UAE delivery teams know well: a permanent hire carrying a finished project, an agency invoice for work that should have been done by an embedded senior in a quarter of the time, an offshore team coordinating around a regulated UAE context they cannot really see. Getting the partner wrong is expensive. Getting the model wrong is more expensive.
This piece is a perspective on how UAE businesses should choose between the four common engagement models. The argument is opinionated. We have a stake, BY BANKS runs the embedded and agency-style models that two of the four labels describe, and the argument is for matching the model to the work, not for steering every decision toward us. We are not arguing that any one of the four is universally best. We are arguing that each has a shape of need it is genuinely best for, and the shape of the need, duration, seniority required, regulatory exposure, and IP ownership, is what should decide the model. A business that decides between models without characterising the need is choosing on habit; a business that characterises the need first picks the right instrument far more often.
The audience for this analysis is founders, CIOs, CTOs, and transformation leads choosing how to resource a specific build, or running a portfolio of builds and noticing the same engagement model being applied to needs of very different shapes. The useful diagnostic question is not "which model should we use" but "what shape is this need, and which of the four is best for that shape, separate from what we usually default to".
The Four Models, Tested Against the Need
Below is a short read. Answer the four factors about a real build you are deciding how to resource, and the four models will surface a strong / possible / poor fit per option, with a brief rationale. The point is not the score; it is that "which model" is downstream of "what shape is the need", and the exercise is built to surface that.
Match the engagement model to the shape of the need
Answer the four factors, then read the fit verdict per model
Why Each Model Is Genuinely Best for Something
The four models exist because they each solve a different problem. An embedded senior engineer is the right answer for a defined piece of senior work that has an end state: a transformation push, a regulatory delivery, a complex integration, a system rebuild, where the seniority is needed for the work and not at the same level once it is delivered, and where the perm hire cycle would lose the window for impact. A software agency or build partner is the right answer for a discrete deliverable where someone outside the client organisation can scope, build, and hand over the whole thing: a new platform, a defined product, an end-to-end engagement with a clear scope and a clear end. A permanent hire is the right answer for an enduring capability where the value comes from continuity and presence, where the seniority is needed every day for years, and where the hiring cycle's cost is small against the relationship's duration. An offshore or distributed team is the right answer when the work is well-defined, the regulation is absent or light, the seniority needed is mid-level execution rather than judgement, and the coordination cost is genuinely accounted for rather than wished away.
Each of those is a strong fit for the shape of need it is matched to and a weak fit for the others. The recurring mistake is treating the question as a preference rather than a fit problem. A business that always uses agencies, or always hires permanently, or always offshores for cost, is applying one instrument to needs of fundamentally different shapes, and the cost of the mismatch is mostly invisible until later, when the perm hire has nothing to do, the agency retainer has drifted into being a de-facto staff arrangement, or the offshore team cannot get a Civil Defence question answered fast enough.
The UAE context sharpens the choice in two specific ways. First, regulated builds (banking, insurance, healthcare, fire safety, government) reward in-region reachability and seniority on the work to a degree that distorts the cost comparison. The offshore option that looks cheaper on the unit rate is often materially more expensive once the coordination, timezone, and supervisory-context overhead is accounted for, and that overhead is hard to forecast accurately before the build starts. Second, the UAE labour market makes senior perm hires slow to source and expensive to carry, which is part of why the embedded-senior model has become more common: it is the instrument designed for the case where senior judgement is needed now and the perm cycle would arrive after the window has closed.
The shift in one observation
"Embedded or agency or perm or offshore" is a question with no general answer and a clear specific one for each need. Businesses that get this right characterise the duration, seniority, regulatory exposure, and IP ownership of the build first, then pick the model the shape demands. Businesses that get it wrong pick the model they default to, then make the need fit. The first reads as decisive; the second is recurring overspend.
Where the Default-Model Approach Breaks
Applying one model to every build breaks in four predictable places.
Perm hire for a finished outcome
Headcount is added for what was really a defined delivery. The build ships, and the business carries the cost and seniority indefinitely, against work that has ended. The perm hire is right for someone, just not for this.
Agency retainer that became staffing
A discrete engagement drifts into an open-ended retainer. The buyer is paying agency rates for what is now effectively augmented staffing, with the agency carrying IP and continuity risk the buyer never agreed to.
Embedded engineer treated as cheap perm
The embedded senior plugged in for a defined outcome stays attached after the work has ended, because no one has the conversation about transition. The model loses its discipline and starts to look expensive against what is now indefinite work.
Offshore on a regulated UAE build
The unit-rate saving disappears in coordination cost and supervisory-context drag. The team is capable; the model is wrong, and the right model would have produced a faster, cleaner, ultimately cheaper outcome.
The Decision in Plain Terms
Side by Side: Which Model for Which Need
| Need shape | Strong fit | Possible | Poor fit |
|---|---|---|---|
| Time-boxed senior delivery | Embedded senior, agency | Perm hire if seniority is genuinely permanent | Offshore for regulated work |
| Enduring core capability | Perm hire | Embedded for transition, then handover | Open-ended agency retainer |
| Regulated UAE build | Embedded senior, in-region agency | Perm hire with regulated experience | Offshore distributed team |
| Defined deliverable, end state clear | Agency, embedded senior | Hybrid embedded plus perm | Perm hire for finite work |
| Commodity execution, no regulation | Offshore, perm mid-level | Agency for delivery push | Embedded senior at senior rates |
The most expensive resourcing mistake on a UAE software build is not picking the wrong partner. It is picking the wrong model, then finding the best partner to run a model that did not fit the work, and discovering the mismatch when the build is already in flight.
What Choosing Well Looks Like
The pattern in businesses that resource builds well is recognisable. The shape of the need is characterised before any model is chosen: how long is this work going to need to exist as something the business runs against, what seniority does it require, how exposed is it to UAE regulatory context, and who needs to own the IP and code at the end. Each model is then evaluated for fit against that shape rather than preference or precedent. Where the need is mixed, an enduring core plus a time-boxed push, which is the common case, the two parts are resourced separately with the right instrument for each rather than forcing one model onto both halves and mispaying for one of them. Cost is compared on the basis of the right model rather than the cheapest unit rate; the cheapest unit rate on the wrong model is the most expensive build.
How This Sits With BY BANKS, Honestly
We have a stake in this argument and it is fair to name it. BY BANKS runs both the embedded senior delivery and the agency-style build-and-hand-over models, the first and second of the four described above. The argument we are making would, if buyers internalised it, send some needs to perm hire or to offshore teams, including needs we would otherwise be commissioned for. We are arguing for matching the model to the work rather than for steering every decision to one of the two models we run. The argument stands regardless of who is chosen for any given build.
The boundary stays clear. We are an independent software engineering company based in the UAE. We design and build software and hand it over. We are not a recruitment agency, we do not place candidates, and we do not provide staffing, payroll, or employment services. We do not act for or on behalf of any UAE authority and we are not affiliated with or endorsed by any authority. On every engagement, the buyer owns its hiring, employment, procurement, contracting, and regulatory decisions and their legal implications. The accountable party leads and owns those obligations; we provide the model that best fits, where we genuinely fit.
Where This Analysis Is Useful
The conversations where this perspective is most useful tend to be at three moments: a business deciding how to resource a specific build and unsure whether the default model is the right one; a leader running a portfolio of builds and noticing the same model applied regardless of shape; or a buyer who has already chosen a model and is feeling the cost of a mismatch. The honest answer is usually the same: each model is genuinely best for a specific shape, the decision is about matching not preference, and the cheapest version of the wrong model is more expensive than the right model at its market rate.
For broader related work, see our perspective on the perm versus contract shift and our perspective on how to choose a software engineering partner. The applied work sits across our operational platforms and technical consultancy capabilities. Get in touch if a 45-minute conversation about a specific resourcing decision would be useful.
Frequently Asked Questions
No. We are an independent software engineering company based in the UAE. We run two of the four models described here: embedded senior delivery and agency-style build-and-hand-over. We do not recruit, place candidates, or provide staffing, payroll, or employment services. On any engagement, the buyer owns its hiring, employment, and workforce decisions and their legal and regulatory implications. We provide the engagement model that fits; the buyer owns the employment relationships.
No. It is an observational reasoning aid to make one point concrete: that the choice between models is downstream of the shape of the need. It is not a methodology, a model, or hiring, employment, procurement, or legal advice, and it describes no specific business or engagement. Real resourcing decisions depend on the specific need, the business, and its own legal and regulatory context, which the buyer owns and should take qualified advice on.
No. Offshore is a strong fit for commodity execution work without regulated context, where timezone and reachability are not decisive. It is a poor fit specifically for regulated UAE builds and senior judgement work, where the coordination and supervisory-context overhead can erase the unit-rate saving. The question is not "is offshore good or bad" but "is offshore good for this build", and that depends on the shape.
Often the right answer for mixed needs. A common shape is an enduring core resourced by perm hires, with embedded senior delivery for a transformation push or regulatory delivery, then handover back to the perm team. Two instruments, deliberately combined, against a need that has two shapes. The mistake is forcing one instrument onto both halves; deliberately combining two is not a hybrid in the woolly sense, it is just the right resourcing for the actual need.
The perm-versus-contract perspective made the structural point that the choice is downstream of whether the need is a permanent capability or a time-boxed outcome. This piece operationalises that point across four common engagement models with the additional factors that decide between them in practice (seniority, regulated exposure, IP). The two are consistent: the question is the shape of the need, and the model follows from the answer.
How to resource a UAE software build is widely treated as a preference question, embedded or agency or perm or offshore, and is in practice a fit question between the shape of the need and the model that best handles it. Each of the four models is genuinely best for a specific shape, duration, seniority required, regulatory exposure, and IP ownership decide the choice, and the most expensive mistake is applying a default model to a need whose shape does not match it. We have a stake in this argument and the argument stands anyway; we run two of the four models and would direct buyers to the other two where those fit better. The boundary is the same as it has been throughout this work: the buyer owns its hiring, employment, procurement, contracting, and regulatory decisions, and a good partner is the one whose engagement model genuinely fits, not the one that runs every engagement the same way regardless of shape.
This article reflects our perspective on resourcing models for UAE software engagements. Figures cited (UAE net employment outlook of +48% for Q3 2025 and 56% of UAE employers planning to increase workforce in Q3 2025) are drawn from the ManpowerGroup Employment Outlook Survey Q3 2025 UAE summary as published; they are point-in-time and the authoritative current data is whatever is published by ManpowerGroup at the time of reading. The four-model fit grid and its rationales are observational and illustrative rather than measured, represent no specific engagement or determination, and are not a methodology, scoring rubric, or hiring, employment, procurement, or legal advice. BY BANKS is an independent software engineering company; we design and build software and hand it over, we run embedded senior delivery and build-and-hand-over engagements, we do not provide recruitment, staffing, payroll, or employment services, we are not a regulated entity in any sector we serve, and we are not affiliated with or endorsed by any authority. On any engagement, the buyer owns its hiring, employment, procurement, contracting, commercial, regulatory, and compliance decisions and responsibility for their implications. This article is not hiring, employment, procurement, regulatory, or legal advice; buyers should obtain qualified advice for their specific circumstances. Public sources used in this piece are listed on our Sources and Data page.
Ready to Build Something?
If this resonated, let's talk about how we can apply these ideas to your business.
Start a Conversation