AI’s silver lining is RegTech – but who will adopt a solution?

Everyone is talking about AI. Every day there’s a new tool, a new partnership, a new “breakthrough”. Same story, different headline.

Tech companies are racing to build the best ‘stack’, lock people into the newest thing, to ultimately help them commercialise it at scale.  That’s the game, we all know it.

We also know all the limitations, the risks: the potential for the slippery slope that exploits the vulnerable, breaches privacy & security, infringes intellectual property rights and takes jobs. This conversation isn’t new.

From a legal perspective, I’m watching what Claude, Clio, and Harvey etc are all doing: the alliances they are forming and the ‘talent’ they are adding to feed their tech stacks. Their intent is to monopolise the market and displace smaller players who can’t compete at that scale, All in order to capture the audience they need to keep training their AI. Yes, it’s a circular narrative, and it appears to be working.

In terms of AI legal learning, lawyers are now being engaged to verify AI advice produced for other lawyers within a CRM system that takes real client information and files as input. The subscription model is straightforward, for now.  Lawyers pay to have a case management system which can produce AI legal advice, which other lawyers outside the CRM system verify so that AI can learn faster.

For Government, it appears they have laid down the gauntlet – all agencies need to find a ‘use case’ and the internal AI wardens must vet (conduct AI impact assessments) and monitor those against the evolving framework of AI Ethics Principles, DTA Policy For Responsible Use, and  the Voluntary AI Safety Standards, amongst other pieces of the regulatory patchwork mentioned in our previous newsletter.

But here’s what I believe has been missed.

AI doesn’t solve much on its own. The gap between AI use cases, collaboration and productivity is something no one is talking about or solving.

AI is fast and the output is impressive. But it’s still just a tool, and it doesn’t ensure the context is captured, provide an audit trail or connect compliant decisions across a system, which is critical for the Commonwealth.

My view is that AI will produce a bottleneck for in-house legal, procurement teams and program areas if not connected with a RegTech solution. It’s already happening. Silos we all know well are being fuelled with content at speed, with no clear path to an outcome due to the time required to review AI-produced content. The ‘human in the loop’ still needs time and good information to achieve compliance and sustainable outcomes.

I see a very clear opportunity here for Government to integrate AI into RegTech and produce a genuine sustainable solution.

In practical terms, this looks like:

1.      Procurement Compliance built in, not bolted on

• Embed Commonwealth Procurement Rule compliance directly into workflows, not after the fact checks

• Introduce real-time flagging of value thresholds, panel requirements and exemptions

• Ensure compliance happens as decisions are made, not after the fact.

2.      Auto-generate audit ready records as part of the design

• Include single source of truth dashboards

• Integrate data and assurance tools

3.      Legislation as code (Rules-as-Code)

• translate legislation and policy to machine readable logic

• introduce decision assurance layers to validate outcomes

4.      Contracts to Data

• Convert contractual obligations into structured, trackable rules

• Dashboard KPI’s and monitor compliance in real time

5.      Grant and Program Compliance

• Tie funding to milestone and condition tracking

• Monitor compliance continuously, not retrospectively

• Flag risks before they become audit issues

6.      Legal connected to compliant outcomes early – no lags

• Connect workflow to collaborative dialogue required with legal in the loop to flag, resolve and help move the project forward.

Sounds straightforward, right?

We all know legacy systems remain a barrier. Investing upstream to save costs later is a difficult task, however, with AI moving so fast, perhaps it’s time to trigger a genuine systems approach. Reducing audit exposure without replacing legacy systems is achievable, and there are sovereign solutions available that international tech companies can’t offer.

The other challenge the major AI platforms haven’t solved yet is how to add the ‘context’ and ‘judgement’, that only humans can provide.

What will the major AI platforms do to address this? Will there be a new shiny AI tech toy to announce and scare the humans with?

I await tomorrows AI headline with excitement.

Until our next newsletter edition – stay curious, empowered by your judgment and ability to add context where it matters.