Transformation Roadmap and Strategic Recommendations

1Executive summary

Deliverable 1 found the gaps and Deliverable 2 set the design: an owned data core, an automated ingestion layer, and specialist tools kept swappable around it. This deliverable turns that design into a plan. It ranks the work by where value, external risk, manual effort and labour cost all point together, sizes it backwards from the hard constraint that Stage 2 must complete before the end of Q4 2026, and lays out how the Group can source it.

The sequence is built to land value early and protect the deadline. The quick wins, AI drafting and an onboarding ingestion pilot, run first and do not depend on any core system going live. The owned data core and the data conversion that feeds it come next, because they gate everything downstream. Onboarding, the heaviest and highest-risk workstream, is the first full proof of the core, and the rest of the priority cluster follows. Everything is grounded in the figures the function heads gave during discovery, not in modelled estimates.

• Onboarding leads, on the evidence. The compliance team spends about 70% of the day on onboarding and periodic reviews, and told us automated onboarding with OCR would address roughly 90% of the team's work. That is the strongest observed case in the Group, and it validates the gap report's ranking.
• The data core gates the deadline. The single client and entity identifier and the golden record are the precondition for onboarding, reporting and compliance workflow. If the conversion slips, Q4 slips. It is the critical path.
• The value is in the labour base. The Group's annual payroll is about AED 36.5m, roughly USD 9.9m, and the priority cluster holds about 72% of it. That is where return is measured, not against the roughly USD 111k of annual licences.
• Hybrid sourcing is the realistic base case. The internal systems team is six people and sits under Finance. It can own bounded work, not the core build against this timeline, so the recommended model has CrossVal build and operate the core while the internal team owns reporting and configuration.

2How to read this report

2.1 How the work was prioritised

The ranking uses four dimensions. Value and external risk are the scores from the gap report's matrix. Manual effort is an evidence-based read on the same one-to-five scale, grounded in the discovery figures rather than measured, so it is calibratable with the function heads. Labour cost and headcount are hard figures from the April payroll file. Where all four point the same way, the work ranks highest. Dependency then reorders within that: anything that reads the master data waits for the data core.

The figures cited against each use case in Section 3 are the ones the function heads gave during the discovery meetings. They are observed statements of current volume and effort, not CrossVal savings claims.

2.2 The closure paths and the four sourcing paths, carried forward

The closure-path tags from the gap report still apply: configure what exists (C), integrate what exists (I), remediate the data (D), build new capability (N). The four sourcing paths from the proposal also carry: commercial software (A), AI tools (B), CrossVal-built and operated (C), internal build (D). Section 3 tags each use case, Section 6 sets the sourcing model.

3Prioritised use cases

The prioritisation map below places every function on all four dimensions at once: manual effort across the horizontal, external risk up the vertical, monthly labour cost as the size of the bubble, and value as the depth of colour. The seven use cases that carry the transformation are the numbered items clustered in the upper right, where effort, risk and cost concentrate together. The lettered, dashed functions sit lower and lighter, and are the deferrals.

Those seven are set out below, ranked by the four-dimension test and then reordered by dependency. Each rationale is grounded in what the function heads observed during discovery.

4The phased roadmap to end Q4 2026

The roadmap is sized backwards from the end-of-Q4-2026 backstop, across roughly six months from the start of Stage 2. It is deliberately staggered so that value lands before the core is complete, and so that the gating item, the data core, has room to run without blocking the quick wins.

4.1 The phases

5Return: the labour base, not the licence line

The return on this programme is not a software saving. The existing licences are roughly USD 111k a year, and reducing them is not where the value is. The value is in the labour base and how much of it the architecture gives back.

5.1 Where the value sits

Group payroll is about AED 36.5m a year, roughly USD 9.9m. The priority cluster, the functions in Section 3, holds about 72% of it. So the roadmap is pointed directly at where the labour cost concentrates. The return is measured as capacity given back in those functions and, in one case, as revenue.

5.2 Three ways the return shows up

• Capacity unlock, in the team's own terms. The compliance team put automated onboarding at roughly 90% of its work. The accounting team set its own goal of moving annual accounts from 8 hours to 4. These are the Group's observed figures, not our projections, and they describe capacity returned to higher-value work rather than headcount removed.
• Revenue leverage, in the outsourced compliance line. The 20-strong outsourced compliance service is billable. Automating its onboarding and monitoring lets each officer carry more client mandates, which is capacity converted into revenue rather than cost. This is the clearest fit to the revenue-per-professional question the sponsor set.
• Growth absorbed without linear hiring. The Group runs about 1,200 entities with heavy inter-entity servicing, where roughly 75% of Rethink's CT revenue comes from M/HQ upsell. Removing the duplicate entry across that servicing chain is what lets the Group grow entities and clients without adding staff in proportion.

5.3 How the levers become margin

Read together, the three levers point at operating margin, the sponsor's own frame of revenue per professional. Margin moves on two sides at once. On the cost side, the priority cluster's roughly 72% share of payroll does more work per dirham once duplicate entry and manual reconciliation come out, so the labour cost of servicing a given book of entities falls. On the revenue side, the billable outsourced compliance line converts freed officer time into additional client mandates, and the drafting and onboarding capacity lets the fee-earning teams take on more work without hiring in proportion.

The inter-entity servicing model is what makes this a margin story rather than a one-off saving. Because a single client is serviced and billed across M/HQ, Rethink and RAA, each duplicate removed compounds along the servicing chain instead of saving once. That is the operating leverage: revenue can grow against a payroll base that grows more slowly, and the gap between the two is margin. The same 1,200-entity book, serviced with less labour and carrying more billable work on the same base, lifts the margin the Group earns on its people, which is the return the licence line was never going to move.

We deliberately do not put a single savings percentage on this. The proposal committed to a plan built on observed data, and the honest position is that the capacity levers above are real and measured, while the conversion of that capacity into a specific financial return is a Group decision about redeployment and pricing, not a CrossVal estimate.

6Sourcing recommendations

The proposal set out three ways to run the build and the steady state. The choice turns on one hard fact from the payroll and org data: the internal systems team is six people, including the Head of Systems, and it reports up through Group Finance rather than holding an independent seat.

The hybrid model is also the lower-lock-in option, because the Group owns the data throughout, so the operating partner can change without the record moving. Two conditions attach to it. The internal systems team needs a standalone technology seat rather than reporting through Finance, which is an operating-model change for Synthesis B. And if the Group wants to internalise the core over time, Stage 2 should carry an explicit hiring and capability plan for that team, sized so the handover is real rather than nominal.

7Dependencies and the critical path

7.1 The critical path

The binding sequence is data remediation, then the owned record and single identifier, then onboarding and everything that reads the master data. The quick wins sit off this path and proceed in parallel. If the data conversion moves, the end date moves, so it is the item to protect above all others.

7.2 The dependencies to clear

AAppendix: assumption log

Continues the logs from Deliverables 1 and 2. Numbering restarts for this deliverable.

End of Deliverable 3. This report sequences and sources the work and frames the return. The Stage 2 go decision rests with Group leadership at the end-of-Stage review.