Insights for the
Modern Enterprise.

Most companies have never comprehensively mapped their business processes. Processes exist in people's heads, in fragmented documentation, and in system configurations that no one fully understands. Process mapping is the act of documenting current state processes in detail: what steps are in the process, who does each step, what systems are involved, how long does each step take, what are decision points and exceptions.

Process mapping is often viewed as a tedious, low-value exercise. In reality, it's one of the highest-ROI investments you can make. Why? Because you cannot optimize what you do not understand. Process mapping creates visibility into how work actually happens. It surfaces inefficiencies and waste that are invisible otherwise. It identifies where processes are broken or non-compliant.

Process mapping ensures that technology implementation aligns with redesigned, optimized processes. Good process mapping has several characteristics. It is comprehensive: it covers all major flows and decision paths, not just the happy path. It is detailed: it documents who does each step, what systems are involved, what information flows between steps. It is accurate: it reflects how work is actually done, not how it should be done in theory.

Seventy percent of AI projects fail to deliver expected value. The common factor? Starting with technology rather than business reality. Many organizations rush to implement Large Language Models (LLMs) or predictive agents on top of processes that are already inefficient, redundant, or fundamentally broken. Automating a bad process only makes it fail faster.

A Business Process Reengineering (BPR) first approach means you simplify and optimize the workflow before you introduce the intelligence layer. We often find that 30% of process steps can be eliminated entirely before AI is even discussed. This 'lean' foundation ensures that when AI is implemented, it is solving the right problems with clean, structured data and clear logic gates.

Consider a recent case where a professional services firm tried to implement AI-driven document review. The project stalled because their document intake process was inconsistent across teams. By reengineering the intake process first, we reduced the complexity for the AI by 60%, leading to a successful rollout that actually saved time instead of creating new technical debt.

For the last 20 years, the build-versus-buy decision has been simple: SaaS is cheaper and faster than building custom software. However, AI-accelerated development has shifted the economics fundamentally. GitHub Copilot and similar tools have reduced software development time and cost dramatically, making custom builds viable for mid-market budgets.

Second, understand the true cost of SaaS: license costs plus implementation, training, support, and customization. A SaaS platform that costs A$100K annually in licenses might cost A$200K annually when you account for all the other costs. Third, compare total cost of ownership over a 3-5 year period. Often you'll find that building custom is actually cheaper.

The difference now is that more problems are worth solving with custom build than was true five years ago. The key is being intentional about the decision rather than defaulting to SaaS because 'that's how we've always done it.' Custom software allows you to own your data, your IP, and your competitive edge without vendor lock-in.

Before you implement AI, you need clarity on whether your organization is ready. AI readiness is not just about having data and infrastructure. It's about having clear processes that you're trying to improve, well-defined success metrics, and organizational alignment on what problems AI should solve.

Process readiness is often the binding constraint. Most organizations have some data and some technology capability, but they lack clarity on what processes should be improved and how AI should fit into those processes. This is where business process reengineering becomes critical.

Fixing the visibility problem manually was the priority; AI scheduling could come later. Once visibility was clear, AI scheduling became feasible and valuable. This discipline prevents wasted investment and dramatically improves success rates across the enterprise.

The Australian Privacy Act underwent significant changes in 2024-2025. These changes raise the bar for customer data protection and increase penalties to A$50M. Mid-market companies are particularly exposed because they typically lack dedicated privacy resources and large-scale compliance teams.

The key changes include: new requirements for explicit customer consent, mandatory breach notification, and enhanced security requirements proportionate to the sensitivity of the data. You need updated mechanisms that clearly explain what data you're collecting and why, along with documented data retention policies.

Companies that are non-compliant face penalties up to A$50M and private lawsuits from customers. The reputational damage from a privacy breach is also significant. Compliance is not optional; it's essential. Mid-market companies should conduct an audit immediately to identify gaps before enforcement intensifies.

Digital transformation projects fail at high rates. Gartner estimates that 60% of initiatives fail to achieve their expected benefits. The common pattern: companies invest in technology without first understanding whether their processes are designed to leverage that technology.

Consider an ERP implementation that was supposed to reduce costs by 20%. They customized the ERP to fit their existing processes, which preserved all the waste and inefficiency of the old system. The company had a shiny new system running their broken processes at scale.

The antidote is starting with business process clarity. Redesign your processes first, then select technology that supports that future state. This approach has dramatically higher success rates and delivers measurable outcomes that actually move the needle on your bottom line.