Why most teams aren’t getting value from AI
By Adrian Randall (pictured), Director & Founder, Arcadian Digital
Across Australia, organisations have rushed to embed Artificial Intelligence (AI) into everyday work. It’s hard to find a software stack these days that doesn’t incorporate some form of “AI”, whether it’s a chatbot, an agent or some other assistant that hooks back into one of the leading large language models (LLM).
The question many business leaders find themselves asking months after deployment, is if those features are converting into productivity gains and driving value. While some AI features might get used for some obvious or novel Q&As, there is a lot of capability being left on the table.
OpenAI has described this shortfall as the “capability overhang”: a growing gap between what AI systems can do and what organisations actually achieve with them. While the tools are powerful, most teams still treat them like interns rather than collaborators.
This is a question of competence, confidence and culture, not access or cost.
From quick wins to complex execution
The first wave of AI felt like a series of quick wins: one‑line prompts for captions, copy tweaks or summaries. That time is fading.
Models such as OpenAI’s GPT‑5.2, Claude’s Opus 4.5 and Google’s Gemini 3 now complete tasks in seconds that once took human specialists half an hour or more. Yet many workplaces still limit AI to simple administrative support or minor edits.
Adopting AI in 2025 was like getting your first laptop. By 2026, it’s closer to being handed an ironman suit. The challenge isn’t owning the suit but learning how to effectively pilot it.
The difference between basic use and advanced capability is the same as knowing how to open a laptop versus knowing how to code on it.
Finding the capability is worth it. Emerging data shows top‑tier AI users extract about seven times more value from the same tools and subscriptions as typical users. That difference rarely comes from better technology; it comes from better methods.
High‑performing teams know how to use structured prompting, chain‑of‑thought reasoning and connect tools across their tech stack. Many have moved towards agentic workflows, where AI systems manage reasoning, planning and execution steps with limited supervision.
If your measures of success still focus on “chat count” or “prompt volume”, the overhang may already be costing you.
Where capabilities go unused and how to close the gap
Two AI capabilities often sit idle: data analysis and reasoning automation. Benchmarks suggest that analytic modules go unused in nearly one fifth of enterprise accounts, while reasoning tools remain untouched in around 14 per cent (OpenAI, 2026).
That’s not a licensing issue; it’s a capability issue. Every unused feature represents lost insight, time or automation potential. Over a year, the unrealised value can surpass the total cost of ownership.
Sadly, teams often purchase enterprise AI subscriptions knowing what’s possible, then fall back to using only what’s familiar.
Closing the capability overhang starts with rethinking how AI integrates into existing operations. It’s not just a helper for surface‑level tasks but a reasoning layer that supports deeper decision‑making.
The organisations seeing measurable returns have stopped bolting AI onto existing processes and started redesigning workflows around what AI can handle end-to-end.
Four practical steps can help:
- Audit current usage. Map each AI tool against its full capability set to find gaps in adoption or overlap.
- Upskill by role and use-case. Replace general AI awareness sessions with training tailored to specific roles, from marketing and operations to analytics.
- Redesign workflows. Turn repeatable processes into AI‑supported workflows that complete multi‑step outputs such as reports, campaigns or insight summaries end to end.
- Establish digital guardrails. Use clear governance and approval paths to maintain data integrity, compliance and trust.
When these elements align, AI shifts from being a support tool to an operational multiplier.
Build capability before the next upgrade arrives
Technology evolves faster than organisational readiness. Teams that bridge their capability overhang now will make decisions faster, operate more efficiently and strengthen their margins.
For others, the gap will continue to widen. The next upgrade will arrive before the last one is mastered. The tools will keep advancing, but outcomes will stagnate unless your people develop the skills to use them properly.
The organisations pulling ahead are those treating AI capability as a workforce development priority, and they are rethinking workflows with their AI capabilities in mind, rather than adding AI to established practices.
The technology is ready. The question is whether your team has the capability to leverage it.
