← case studies
Published
May 2026

How a Care Payroll Firm Collapsed 35 Systems to One

A disability-care ops team was drowning in 400,000 files across 35+ systems. An AI-built SharePoint hub now holds the lot — live in 90 days at 90% ROI.
90%
Recouped at project go-live
4–8 weeks
Implementation Time
$25K – $100K
Project Cost

the challenge

Carer Solutions is a Melbourne-based provider of payroll, compliance, and operational services to the disability and community care sector. They navigate SCHADS Award complexity, plan management, and rostering so their clients don’t have to. Their people are exceptional. Their systems were not. 400,000+ files scattered across SharePoint and OneDrive with no structure 19 levels of nested folders navigated purely by muscle memory An expensive project management tool (Wrike) everyone hated but couldn’t escape 35+ disconnected systems with zero documentation “Everyone sees everything” permission model — no role-based access NDIS audit looming with unknown compliance exposure

what they built

Rather than spending 6–12 months cleaning up 400,000 files, we built a completely new digital environment and migrated only what mattered. What we delivered SharePoint Intranet — A central hub with 10 interconnected sites aligned to how they actually work. Staff walked in and started using it immediately — no training backlog, no confusion. Knowledge Management System — Single source of truth for policies, processes, and procedures. Version control and review cadence built in from day one. Compliance Centre — Seven compliance registers (3 active + 4 historical) with full audit trails preserved from legacy systems. Audit-ready immediately. Wrike Decommissioning — Extracted two years of compliance knowledge from tickets. Turned that history into structured FAQs and SOPs. Eliminated the annual licence cost entirely. Salesforce Audit — Analysed 35 licences and identified significant annual savings through licence optimisation. Security Overhaul — From “everyone sees everything” to proper role-based access control with 20 Entra ID security groups. Participant data properly protected. AI compressed the discovery and synthesis layer that normally consumes 6–12 months of consulting. We pointed it at 412,000 files, 1.7M permission grants, and 6,154 Wrike tasks — and got a categorised inventory, PII risk map, and behavioural picture of the organisation back in hours, not weeks. That meant we could build the new SharePoint architecture from their actual operating model — not a generic template — and extract their knowledge management content directly from 151 historical Wrike tickets rather than running months of workshops to draft it from scratch. The specific change: AI provided the domain depth that traditional consulting buys through extensive client interviews. ~35 client meetings replaced the usual 50+, backed by 175 internal sessions doing the synthesis. The client saw polished, fully-furnished deliverables — not drafts, not frameworks needing months of internal implementation. That's why 90% of the investment came back at go-live: staff walked into the new system on day one and used it, with no training backlog or "getting used to it" period eating the benefits.
Real Minds AI ran the engagement as a tight 12-week build rather than a multi-month cleanup. The first three weeks were discovery and assessment: the team mapped all 35-plus disconnected systems and pointed AI at the full estate — roughly 412,000 files, 1.7 million permission grants, and 6,154 Wrike tasks — to produce a categorised content inventory, a PII risk map, and a behavioural picture of how the organisation actually worked. That diagnosis, compressed from the usual six-to-twelve months into hours, let the team design the new environment around Carer Solutions' real operating model instead of a generic template. Weeks four through eight were architecture and build: a ten-site SharePoint intranet, seven compliance registers, and a role-based security model using twenty Entra ID groups. Weeks nine through eleven covered migration and training — moving only the content that mattered, training six teams, decommissioning Wrike, and extracting two years of compliance knowledge from 151 historical tickets into structured FAQs and SOPs. The system went live in week twelve, with staff productive from day one.

best fit for

Small to medium enterprises drowning in disorganised systems and content
Ai ROLE
The AI ingested ~412,000 files and 1.7 million permission grants from SharePoint and OneDrive to produce a categorised content inventory, a PII risk map, and a behavioural picture of how the organisation actually operated. It analysed 6,154 Wrike tasks and extracted compliance knowledge from 151 historical tickets, turning two years of operational history into structured FAQs and SOPs. Humans made every architecture and migration decision; the AI provided the diagnostic substrate and content-extraction layer.

impact

90 days with 90% ROI achieved on go-live

Speed creates value 90 days maintains momentum. Teams jump to the new system while energy is still high. Delay kills adoption.

80% less internal implementation work

Turnkey delivery Staff walked into a finished system. No construction debris. No months of internal implementation. No training backlog.

90% increase in compliance case rates meeting KPIs

AI processed 400,000 files and extracted patterns. Humans made every decision that mattered. Instead of just moving data, we analysed patterns. Two years of compliance tickets became structured FAQ content and processes.

implementation complexity

Weeks 1–3 Discovery & Assessment Mapped all 35+ disconnected systems, analysed 400,000 files, identified what mattered. Weeks 4–8 Architecture & Build Designed 10-site SharePoint architecture. Built compliance registers. Created security model with 20 Entra ID groups. Weeks 9–11 Migration & Training Migrated critical data. Trained 6 teams. Decommissioned Wrike. Extracted two years of compliance knowledge. Week 12 Go-Live Staff walked into a finished system. No construction debris. No months of internal implementation. Done.

Tracy Anthony

Founder & CEO
Real Minds, Artificial Intelligence
Founder of Real Minds AI, advising enterprise public agencies, NFPs, and mission-led SMEs on AI adoption and operating models that hold up after the project team leaves.
Get an intro
Talk to this team
industry
HR & People Operations
business organization
Operations
AI TYpe
Generative Design & Content
value type
Risk & Compliance

more from this expert

No items found.

frequently asked questions

How did a small care-sector firm collapse 35 disconnected systems into one?

The experts ran a tight 12-week build instead of a multi-month cleanup. They pointed AI at the full estate — roughly 412,000 files, 1.7 million permission grants, and thousands of tasks — to produce a categorized content inventory, a PII risk map, and a behavioral picture of how the organization actually worked, then designed a new SharePoint environment around that real operating model and migrated only what mattered. The system went live in week twelve with staff productive from day one.

What AI tools and approach were used to consolidate the systems?

The work centered on generative design and content, built on Claude. AI processed roughly 400,000 files to extract patterns and categorize content, mapped PII risk, and turned two years of historical compliance tickets into structured FAQs and SOPs, while humans made every decision that mattered.

What results did the care-sector firm achieve?

The build delivered an estimated 80% less internal implementation work, reached 90% ROI on go-live within 90 days, and drove a 90% increase in compliance case rates meeting KPIs — collapsing 35-plus systems into a single SharePoint environment with role-based security.

How long did the consolidation take?

A four-to-eight-week range in practice — run as a tight 12-week build, with discovery in weeks one to three, architecture and build through week eight, and go-live in week twelve.

Who is this approach best for?

Small to medium enterprises drowning in disorganized systems and content that need a fast, AI-assisted rebuild rather than a slow cleanup.

Have a similar challenge?

Ask whether this would work for you, or describe what you're trying to solve.
TELL US WHAT YOU'RE EXPLORING