Advisory · Commercial leadership
Commercial AI strategy, written as a commercial plan.
An AI strategy for marketing, sales and commercial operations that reads as a commercial plan with AI inside it, rather than a technology plan appended to one. Sequenced over ninety days, with a named owner and a number against every move.
It also establishes what not to fund, which in most organisations proves to be the more valuable half of the document. Enterprise AI investment in the UAE rose by 105% in a single year and 95% of that expenditure has produced no measurable return, and the difference between those two figures is very largely a question of sequencing.
UAE enterprise AI maturity research 2026 · MIT 2025
The method
Four questions, asked of each commercial function in turn.
The analysis is deliberately bounded. It examines the functions in scope rather than the whole organisation, which is why it concludes in weeks rather than months and why the resulting document is short enough to be acted upon.
-
01
Value - where is the money actually going?
Which activities inside the function consume the most cost, and how much of that cost sits in work that is high in volume, repetitive and already measured. This is ordinarily the point at which an organisation discovers that its largest recoverable expense is not the one it had assumed, and that the assumption was formed before the cost of producing that work changed.
-
02
Friction - what is slow, expensive or simply not happening?
The work that takes a week when it should take a day, the work that is rebilled by an agency, and the work that has quietly been abandoned because there was never time for it. The third category is the one most frequently overlooked, and it is where the largest commercial opportunities tend to be found, because nobody is measuring something that is not being done.
-
03
Data - what may be used, and under what conditions?
Which categories of client, pricing and contractual information may be placed into which systems, where that data physically resides, and what your obligations are under the United Arab Emirates data-protection framework and any contractual undertakings you have given customers. This question is answered before anything is implemented rather than afterwards, because reversing a decision in this area is expensive and occasionally public.
-
04
Asymmetry - where does this create an advantage rather than parity?
Automating work that every competitor is also automating produces no advantage, since everyone acquires the same capability at the same price in the same quarter. The question that matters is which changes are difficult for others to copy, and those are almost always the ones that depend on something the organisation already possesses: its data, its relationships, or the particular way it goes to market.
The deliverable
A ninety-day plan, and a list of what not to fund.
- A sequenced ninety-day plan, with each move carrying a named owner inside your organisation and a number against it, ordered so that the items funding the later stages come first.
- A position on every commercial function in scope, placed across the four levels at which AI can be applied, with the two or three widest gaps identified.
- An agency and outsourcing review: which work should return in-house, which continues to justify a premium, and what the difference is worth annually.
- A written data and governance position, covering what may be placed into which systems and what your UAE obligations require, in language a commercial team can apply.
- A baseline, taken before anything changes, so that the result can later be argued from your own figures rather than from ours.
- An explicit list of what not to fund, which is the section most often used in the following board meeting.
Where this sits
The strategy is the first of five stages, and each is scoped separately.
Organisations frequently engage only the first stage and proceed no further with us, which is a legitimate outcome and occasionally the correct one. Each stage is priced and agreed on its own terms rather than as a committed programme.
-
01
Baseline the function
Hours spent on the target workflows, cost per output, outsourced expenditure, and how AI is already being used informally by the team, which is almost always more than the organisation believes.
-
02
Strategy with the leadership
The commercial plan is agreed and funded before a single cohort runs. This order is not negotiable, because a trained team returning to an unchanged week absorbs the training without trace.
-
03
-
04
Rebuild the top workflows
The two or three carrying the most cost, rebuilt end to end and documented as they are built, so that they survive the departure of whoever was closest to them. Where something genuinely needs to be built, our own developers build it.
-
05
Re-measure, report, and leave
Against the original baseline, in your own numbers, written up in a form that can be taken to a board. Then a documented handover, and our departure. An arrangement that depends on the consultant remaining has transferred nothing.
Suitability
Who this is for
Organisations of roughly fifty people upwards with a commercial function large enough that the work inside it is measurable: a marketing team, a sales team, or a commercial operations group whose output can be counted. It is written for the person accountable for that function's cost and revenue - a chief commercial officer, a managing director, a marketing or sales director, or a general manager.
It suits organisations that already hold AI licences and cannot demonstrate what they have produced, which in this market is the substantial majority.
Who it is not for
Organisations seeking a technology strategy rather than a commercial one. Where the question is which model to deploy, how to structure a data platform, or how to build machine-learning capability, a technical consultancy is the correct choice and we will say so in the first conversation rather than the third.
It is also unsuitable where the commercial leadership is not prepared to change how the work is organised. The analysis will produce the same findings regardless, but a plan that nobody is authorised to act upon is an expensive document.
The order matters
Training a team before leadership has agreed what changes is absorbed without trace.
A trained team returning to unchanged targets, unchanged reporting and an unchanged working week will revert within a quarter, and the organisation will conclude that the technology did not work. The strategy is therefore agreed and funded before a single cohort runs. This is the one part of the sequence that is not negotiable.
Common questions
What leadership teams ask in the first conversation.
How is this different from an AI strategy from a technology consultancy?
A technology consultancy begins with the technology and reasons towards the business, which produces a capable architecture and a long list of things that could be built. We begin with where the cost sits in your commercial function and reason towards the technology, which produces a considerably shorter recommendation and a substantially cheaper one, because most of what could be built should not be. The practice was founded by a former Chief Commercial Officer rather than by a technologist, and that sequence is the whole of the difference.
How long does the strategy take to produce?
Weeks rather than months, because the analysis is bounded by the commercial functions in scope rather than by the size of the organisation. Discovery interviews and the baseline occupy the early part and are conducted by video call. The leadership session itself is a working day with up to twelve executives, conducted online, which allows participants to join from more than one office and keeps the group working against its own systems and figures throughout. The plan is then presented to you and sent across in full, together with the baseline figures, the governance position and everything else needed to act on it. The plan it produces is written against a ninety-day horizon, which is the period over which the moves it contains are sequenced.
How is an engagement priced?
Engagements are priced per programme rather than by the hour or by the number of participants, because what an organisation pays should follow the scope of the work. Three factors determine the figure: how many commercial functions are in scope, whether the engagement is advisory alone or continues into training and rebuilding the workflows afterwards, and whether anything genuinely needs to be built. A strategy engagement for a single function therefore sits at one end of the range and a multi-cohort programme with a development component at the other. We give you a figure during the first conversation rather than across several meetings.
Do we have to continue into training and implementation?
No. Each of the five stages is scoped, priced and agreed separately, and a number of organisations engage the strategy alone and carry out the remainder with their own people, which is a perfectly reasonable decision where the internal capability exists. The plan is written so that this is possible: every item carries an owner inside your organisation rather than inside ours, and the document does not assume our continued involvement in order to be usable.
Are you tied to any particular AI vendor?
No. There are no vendor relationships, no referral arrangements and no commission on anything recommended, which matters because a considerable proportion of the advice available in this market is provided by organisations that sell or resell the systems they recommend. Where existing tools can be configured to do what is required we will say so, since a system that was not needed is the most expensive category of expenditure an organisation can incur in this field.
VAIY