The Hidden Cost of "Good Enough" AI: Why Law Firms Need to Look Beyond Their PMS
Most Australian law firms are approaching AI adoption through the path of least resistance: the AI features bolted onto the practice management system (PMS) they already pay for. It's an understandable instinct. The integration is native, the rollout is low-friction, and it feels like a natural extension of a tool the firm already trusts.
But "convenient" and "future-proof" are not the same thing, and firms need to ask a harder question before they settle: what was this AI actually built to optimise for?
Built for the Back Office, Not the Practice of Law
Most PMS platforms were designed first and foremost as accounting, billing, CRM and reporting systems. That's their DNA, and it's where their AI investment naturally goes: smarter time capture, automated invoicing, cleaner trust accounting, better matter reporting. These are genuinely valuable capabilities, and no firm should dismiss them.
But legal work itself, drafting, contract review, legal research, document analysis, chronology building, is a fundamentally different problem. It requires reasoning over dense, unstructured legal content, applying jurisdiction-specific law correctly, and producing outputs a lawyer can actually rely on. That is not a natural extension of an accounting engine. It's a different discipline entirely, built by different teams solving a different problem.
When a PMS vendor adds "AI" to its product, the priority is almost always to protect and extend the core system of record, not to build best-in-class legal reasoning. That's not a criticism, it's simply where their commercial incentives sit.
The Integration Trap
The pitch usually sounds reassuring: "it already integrates with your PMS." But look closely at what that integration actually covers. In most cases, it extends to matter data, time entries, billing, and basic document storage, the operational layer, not the intellectual work of legal practice.
Ask what happens when your firm wants to:
- Run a sophisticated agentic workflow across a specific document set for a niche practice area
- Apply a jurisdiction-specific legal reasoning model that a generalist PMS vendor has little incentive to build
- Adopt best-in-class AI research or drafting capability as the market moves, rather than waiting for your PMS vendor's roadmap
If the honest answer is "we'd have to wait for our PMS vendor to build that, if they ever do," your firm isn't future-proofed. It's locked into a single vendor's pace of innovation, in a category where the pace of innovation is now measured in months, not years.
The Real Question to Ask
This isn't an argument against PMS-native AI, for accounting, reporting and CRM tasks, it's often the right, low-friction choice. The real question is whether your firm has confused "integrated" with "capable," and whether the tools handling your highest-value work, the legal reasoning itself, are being built by people whose core focus is legal practice, or by people whose core focus is your ledger.
Firms that get this distinction right won't just adopt AI faster. They'll adopt the right AI for the right job, and retain the flexibility to bring in best-in-class legal-specific tools as the market evolves, rather than being tied to whatever their PMS vendor decides to prioritise next.
Future-proofing isn't about having AI. It's about making sure the AI doing your firm's actual legal work was built for that job in the first place.