AI Won’t Save You Money – Unless You Know Where to Look

The promise of artificial intelligence in government is compelling: faster services, smarter decisions, and lower costs. Unfortunately, a lot of these benefits are broadly distributed, and it’s pretty hard to harvest 13 minutes a week from each employee. To realise financial benefits agencies must deliberately harvest from specific use cases, and use these savings to fund broad AI adoption.

Think of the use cases we keep hearing about: drafting emails, summarising meetings, reviewing long documents, faster search and retrieval. Where can we find teams with a high portion of their time spent on these tasks? What about the secretariats that are scattered around every agency – if we can find 2-3 APS level secretariat staff to redeploy and eliminate those positions, we’ve suddenly paid for Copilot or GovAI for at least the whole Division, if not the Group.

Here are some other ideas common across Government:

• Policy teams: preparing for, summarising and responding to public consultations.
• Parliamentary and executive teams: monitoring Government proceedings (e.g. Senate estimates) for relevant topics, and summarising
• Legal teams: precise search and retrieval of previous advice, cases etc. (save from engaging external legal advice where it’s been covered before)
• Comms and change teams: High-speed first-drafts of comms materials, user guides etc.

All of these can be implemented with relatively little investment, but must be done intentionally or there will be no measurable benefits to harvest. Deliberately centralising these common use cases to a smaller AI-enabled team will allow consistent quality controls to be applied and give agencies the best chance of harvesting meaningful savings. Finding just one FTE from each of the above cases creates a lot of financial room for investing in the more sophisticated AI applications, or returning to the bottom line budget.

We’re heading into a year in the federal context where budgets are tight for almost every agency, and scrutiny is high. By tracking metrics, validating outcomes, and ensuring that gains are visible and reusable, agencies can demonstrate to Government genuine innovation being applied to rein in spending. The agencies that succeed won’t be the ones with the most pilots, but the ones that turn pilots into practice, and practice into measurable impact.

If you want to talk about how to implement AI in practice, let’s chat.