Almost every conversation about AI eventually comes back to productivity.
Business owners see reports generated in seconds, emails drafted instantly, and routine admin automated. With research suggesting employees can complete some tasks up to 30-50% faster using generative AI, it’s easy to assume cost savings will naturally follow.
Sometimes they do.
If someone spends less time on repetitive work, they have more capacity for higher-value activities. That is one reason AI adoption has accelerated so quickly, with Microsoft and LinkedIn’s 2024 Work Trend Index finding that 75% of knowledge workers already use AI at work.
What I’ve noticed, though, is that the financial return can be harder to measure than many expect.
Take proposal writing. AI can produce a first draft in a minute rather than half an hour. But proposals still need review. Facts need checking. Numbers require validation. Language often needs refining to match the audience.
The final result is usually stronger because someone has improved what the AI produced.
That is not a criticism of AI. It’s simply how useful tools work.
A spreadsheet doesn’t replace accounting knowledge. A power tool doesn’t replace a skilled tradesperson. In the same way, AI often amplifies expertise rather than replacing it.
Another challenge appears once AI starts spreading across a business.
Marketing adopts one tool for content creation. Operations uses another for research. Sales finds an assistant for meeting notes. Before long, subscriptions begin to stack up.
Industry studies suggest many organisations now use dozens of AI-enabled applications, often with significant overlap in capabilities. As a result, businesses can end up paying for several tools that solve similar problems.
I’ve also seen situations where employees create workarounds when approved tools don’t meet their needs.
A personal account gets used. A free AI platform fills a gap. Information starts moving between systems that were never designed to work together.
At that point, the issue goes beyond cost. It becomes a question of security, compliance, and visibility. In fact, Microsoft’s Work Trend Index found that many employees are bringing their own AI tools into work without formal oversight, creating challenges for governance and data protection.
To get the strongest results from AI, spend less time asking which tool to buy and more time identifying where work is slowing down.
- Are customer enquiries taking too long to handle?
- Are reports consuming hours of manual effort every week?
- Is someone spending their day moving information between systems?
When AI is applied to a clearly defined business problem, it’s much easier to measure whether it’s delivering value.
The opposite happens when AI is introduced because everyone else seems to be using it.
The technology arrives first and the business case gets worked out afterwards.
That approach can still produce useful outcomes, but it makes it far harder to judge whether the investment is paying off.
The encouraging thing is that most organisations are still early in this journey. Research from McKinsey shows that while AI adoption is increasing rapidly, many businesses are still working out how to translate experimentation into measurable business results.
Now is a good time to step back and assess what’s changed.
- Are people completing work more efficiently?
- Have customer experiences improved?
- Is the business operating more smoothly than it was six months ago?
Those questions will reveal far more than a software usage report ever will.
If you’d like help reviewing where AI is helping, where it’s creating extra work, and whether it’s delivering genuine business value, get in touch.
Main improvements
- Added recognised industry statistics to strengthen credibility.
- Replaced generic statements with measurable business examples.
- Expanded the “hidden cost” and “shadow AI” sections to reflect real-world governance concerns.
- Kept the same conversational style while making each point more substantial and persuasive.