Scope and phasing after the Fusion plan

What the Fusion plan says about Catalyst Platform

Source: OCLA Fusion Strategy and Operating Plan, presented 15 October 2026. The plan describes a target, and says its capabilities, budgets and rollout still need confirmation.

Possible scope changes

Compared with the agreed Phase 1 plan and the changes already recorded.

Reporting across all practices and SMMEs

Catalyst and OCLA want reports across the whole network. Each practice's data stays secure and managed by that practice; the network view may be anonymised.

• Today the platform prevents this, on purpose. Every query is confined to one practice, and nothing joins one practice's clients to another's. That stays true for identifiable data.
• Proposed: a separate, de-identified copy for analysis, kept in South Africa and away from the live system. Each night it would receive figures, ratios and signals per business and month, with sector, size, province and entity type, and a coded identifier so a business can be followed over time. No names.
• Two levels. First, results in groups only, each group drawn from enough businesses across enough practices that no practice's clients show through. That level is anonymous. Second, if wanted later, records under the coded identifier for research. Under POPIA those still count as personal information, because the code can be reversed, so they need the same care as named data.

Before any data is used this way: the practice agreement must allow de-identified, combined use of client data for Catalyst's own purposes. The clause should be in the first practice's agreement, because data collected without it is hard to use later. Xero's terms also need checking: they already forbid training AI models on Xero data. Practices outside the OCLA group may have commercial concerns too.

400 practices and 200,000 SMMEs

Estimates, to be checked against the pilot's real data.

Is the technology right? Yes. The Cape Town design (Cloudflare at the edge; the database, files and keys in Cape Town) suits this load, because the heavy work is reading from Xero and storing data, and the design keeps both in South Africa. It needs the planned scaling work brought forward, the separate analysis copy, and moving the rest of prod from the EU to Cape Town before the second practice, which also settles where client data is kept.

Xero certification

Xero limits how many businesses one app may connect to. The limits as published by Xero, last checked August 2026:

• Connecting each business by hand doesn't avoid certification. Xero counts every business connected to the app, however it was connected.
• The limit is for the whole app, not each practice. The pilot (up to 50 businesses) needs no certification. The second practice will most likely take the total past 50.
• A separate Xero app per practice isn't a way round it. It would mean hundreds of apps to secure and run, and Xero would most likely see it as getting around its tiers.
• Start certification during Phase 1. Xero publishes no timeline for it. It doesn't hold up the pilot.

Proposed phasing

Nothing built in Phase 1 stands in the way of these. The analysis copy reads the same tidied data, providers are already configuration, and the scaling work is planned.

Decisions we need from Catalyst

1. Is the first release proven by the full client case, including fees and income reconciled?
2. Are connected individuals in scope, and when?
3. Network reporting: results in groups only at first, or research records as well? Who drafts and signs the practice agreement clause?
4. When do the 400 practices join? This decides when Xero's higher tiers and the move to Cape Town must be ready.
5. One web address for every practice, or one per practice?