Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Businesses should start AI strategy by diagnosing and prioritising organisational problems—not by asking what tasks they can automate. In an opinion article published on 5 October 2026, Martin Paton, CTO at CreateFuture, argues for identifying the problems that matter, then testing AI in one focused, measurable, executive-sponsored use case.
Why starting with automation can lead businesses astray
“What can AI automate?” sounds practical, but it begins with a technology capability rather than a business need. That can direct attention toward routine tasks even when more consequential problems—such as fragile systems, duplicated technology, compliance concerns, or avoidable cloud spending—remain unexamined.
Paton’s argument is that AI can be useful as a diagnostic aid before it is used to automate work. It may help an organisation investigate technical debt, continuity risks, architectural weaknesses, compliance issues, and unused cloud spend. The goal is to understand where the organisation is exposed or wasting effort, then decide which issue deserves attention first.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problems
Paton also describes a client case in which an investigation surfaced substantial unused cloud spend. This is his account, not an independently verified result, and he gives no exact amount. It illustrates the proposed diagnostic approach; it does not establish that AI will uncover savings in every organisation.
Choose the problem before choosing the AI
The first decision is not which model or product to buy. It is which business objective needs support, and how the proposed work would advance it. Paton frames the core questions this way: “Success, therefore, isn’t defined by the tech you buy, but by the questions you ask – and these should start with two fundamentals: which core organisational challenges do we need to solve, and in what order?”
#1 Best Overall
That ordering matters. A list of possible AI applications is not a strategy unless it connects to a prioritised business problem. A useful initial assessment can look for issues that affect important operations, expose the organisation to risk, or consume resources without adequate value. The assessment should then produce a short list of problems to investigate—not an automatic decision to apply AI to all of them.
How to run a focused AI initiative
Paton recommends a thin-slice approach: select one high-impact use case, secure executive sponsorship, define measures of success, and progress iteratively toward a longer-term objective. A practical sequence is:
- State the business problem. Describe the operational or organisational issue in terms decision-makers recognise, rather than naming an AI feature as the goal.
- Set a baseline and an outcome. Identify what is happening now and what measurable change would count as improvement. Faster output alone may not demonstrate business impact.
- Choose one bounded use case. Limit the initial effort to a focused area where the organisation can assess results and manage the consequences of errors.
- Assign an executive sponsor. Give a senior owner responsibility for connecting the work to the business objective and resolving decisions that cross team boundaries.
- Review results and adapt. Compare outcomes with the baseline, learn from the implementation, and decide whether to refine, expand, or stop the effort.
This is different from broad experimentation without a clear endpoint. The point of a small initial scope is not that every AI initiative must stay small; it is that expansion should follow evidence and a defined roadmap rather than enthusiasm alone.
Measure business impact, not just activity
Paton discusses a 95% figure attributed to a 2025 MIT report, describing it as pilots that failed to show direct P&L impact. He cautions against treating that as a universal verdict on AI. The underlying report’s denominator, methodology, and exact outcome definition are not established in the source account, so the number should not be read as a verified failure rate for all AI projects or as proof that AI technology does not work.
Rank #3
The more useful lesson for a business planning its own initiative is to define what it intends to improve and how it will judge that improvement. A pilot can produce activity or a working demonstration without showing a meaningful business result. Clear measures make it possible to distinguish those outcomes.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.Include cost and risk in the plan
Paton notes that enterprise AI may use consumption-based pricing, where usage and processing volume can affect costs. That is a reason to include expected usage and ongoing operating costs in planning, not a universal description of every provider’s pricing. No specific rate or savings figure follows from his observation.
For risk management, the NIST AI Risk Management Framework organises work into four functions: Govern, Map, Measure, and Manage. NIST’s Generative AI Profile is a companion resource for considering risks in common cross-sector activities, including large language model use, cloud-based services, and acquisition. These are voluntary risk-management resources; they do not guarantee a return on investment or endorse a particular adoption strategy. Read the NIST AI Risk Management Framework and Generative AI Profile information.
Quick Recap
Best Value
What to take from Paton’s argument
The headline is a provocative opinion, not an empirical finding about every business. Paton’s central recommendation is narrower and practical: diagnose the organisation’s challenges, decide which ones matter most, and only then select an appropriate AI use case. A focused, sponsored initiative with measurable outcomes gives leaders a basis for deciding what to do next.
Read Martin Paton’s article in The AI Journal.
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
