Feedback and responses
All the feedback we received, in one place, labelled by source, with what we did about each point. The learning plan is at the end.
Third-party review independent external review
An independent external review of the Stage 1 deliverables, obtained via Greg. The points are summarised into themes here; a full item-by-item response also exists.
| What was raised | How we've responded |
|---|---|
| Quantify the case. Findings were directional; inefficiencies, ROI, benefits and value realisation were not measured. | A bottom-up, tiered cost-savings model now sizes savings per workflow from loaded team cost, with a conservative floor that still pays for the programme in its first year. A benefits register gives each target a baseline, an owner and a KPI, measured at the WS0 kickoff. See the running-costs note and the savings model. |
| Prove build-versus-buy; keep it vendor-agnostic. The owned core was not rigorously compared with alternatives; data ownership was conflated with custom build; sourcing was not independent. | The requirement (data ownership and diligence-grade quality) is separated from the build decision. WS0 evaluates commercial platforms at equal depth, selection authority sits with the Group, and there is an exit ramp at every stage. Ownership is guaranteed whichever route is chosen. |
| Lead with process, operating model and adoption. Too internally and technologically focused; end-to-end process, client experience and change management under-developed. | The plan is reframed efficiency-first, opens with a decisions-required summary, and carries a full enablement track (the learning plan below). The operating model was resolved at the 15 July planning session, and client-experience improvements are named. |
| Treat AI as a governed capability. AI was positioned mainly for extraction and drafting. | AI sits as a governed enabler layer: per-function prompt libraries with a usage and security policy, and a maturity path from drafting toward decision support. The platform choice stays open until the libraries are proven in use. |
| Make it navigable and decision-first. Text-heavy, repetitive and visually weak; the executive so-what was unclear. | The deliverables now run as an interactive portal, with the executive summary and the visuals up front, this review pack, and a decisions-required summary for each document. |
| Firm up the timeline and gates. The timeline looked optimistic; decision gates were undefined. | The close was moved to end February 2027 to take risk out of the schedule, and delivery now runs as three gated releases with acceptance criteria and payment tied to each gate. |
| Define the ambition. No clear transformation ambition or measure of success. | Agreed with the Group on 20 July: efficiency at scale is the goal, enterprise value the outcome, client experience retained or improved; success is measured as freed capacity against the baseline-and-target table. |
Simon's review architecture
Simon's questions were architectural rather than commercial: how the system behaves, and who owns it.
| What was raised | How we've responded |
|---|---|
| Which system wins on a contradiction? | A single contradiction rule (Appendix C): data is set one-way at onboarding, the core stays master, every change carries a timestamped audit entry, and protected fields pass an approval gate before they move. |
| How does finance stay consistent when the ledger lives in Xero? | Internal finance is kept per entity in Xero; the core draws only the relevant subset (invoices, amounts due, revenue by client, cost of sale) and consolidates across entities. The full ledger stays local. See the field map. |
| Post-close corrections without manual overhead? | Minor changes are absorbed automatically; material changes route to an approval gate with a full audit trail, so nothing has to be re-keyed across systems. |
| Own it, don't rent it. | Ownership by design: the schema, data and IP are the Group's on an open, portable stack. CrossVal builds and operates, the Group owns the outcome, and can change operating partner without the record moving. |
Greg's steers commercial and delivery
| What was raised | How we've responded |
|---|---|
| Efficiency first. Efficiency the goal, enterprise value the outcome, client experience retained or improved. | This is now the headline framing across the proposal and the executive summary; we never lead with enterprise value. |
| One proposal, two prices. | A fixed build to the end-February close, then a month-to-month asset-building arrangement after it. See the proposal and the running-costs note. |
| Build assets, not support; capitalise it; no senior internal hire. | The work is structured to support capitalisation as asset-building. CrossVal carries the senior capability through a part-time embedded lead; any Group hire is a more junior role, upskilled through the super-user track. |
| Detail the training and pull the team in early. | See the learning plan below: three tiers of training that include everyone from week one. |
Serena's input delivery and adoption
Serena's input came through the planning sessions: how the plan lands inside the team that has to live with it.
| What was raised | How we've responded |
|---|---|
| Adoption has to be run from inside the Group. | Serena runs the adoption programme as project manager: she coordinates the champions network and owns completion tracking, reported at each release gate. |
| The plan has to respect the team's real bandwidth. | The heavier Group-side work sits after 1 October, quick wins are kept light on Group effort, and foundation training is short, self-paced work the team can absorb alongside the day job. |
| Bring people in early, before go-live. | Enablement starts in week one with the company-wide launch and the foundation training, so no one meets the new system for the first time at go-live. |
Anna's Stage 1 feedback domain accuracy
Corrections to how the deliverables described the Group's operations, all applied in the 14 July revision.
| What was raised | How we've responded |
|---|---|
| Origination and onboarding. Client origination starts at M/HQ, and each entity onboards its own clients separately. | Corrected across the deliverables; no central-onboarding claim remains. |
| Onboarding ownership. Onboarding is owned by each entity's MLRO; the gap is the absence of a single consistent process, not the absence of an owner. | Reworded throughout to describe the process gap, not an ownership gap. |
| System and screening specifics. The record systems, and that screening runs on the internal client base regularly, not only at intake. | Corrected, with the R&C data segregation stated consistently, and carried into the field map. |
Learning plan enablement
In response to Greg's steer, training starts on day one and runs through the whole programme instead of arriving at the end. Serena tracks completion and reports it at each release gate.
Tier 1 · Foundation, for everyone, from week 1
General familiarity, assigned to all
A short, self-paced set of language-model literacy courses assigned across the Group, kept platform-neutral, with a fixed completion window (target three weeks) tracked per person. It is kept light on purpose so it can run during the constrained period before 1 October.
Company-wide introduction
One launch for the programme and the transformation
One session that launches the training programme and the transformation together, so everyone hears what is changing, and why, before it reaches their desks.
Tier 2 · Super-users
A champion per function
A named champion per core function (finance, compliance, corporate services, tax, onboarding), given the deeper hands-on training and ownership of that function's prompt library. They become the in-house experts once CrossVal steps back.
Deepening, from week 12 as workflows go live
Role-based, hands-on, measured
Led by the dedicated Training and Education Specialist: role-based curricula per function, hands-on workshops for each new workflow as it lands, train-the-trainer for the internal systems team and the champions, quick-reference guides, with completion tracked per department.