Many organisations begin their AI journey by choosing tools. But tools alone rarely transform the business. In the next article in our series on preparing for the AI shift, Patrick Walsh, our Chief Commercial Officer, looks at why leaders should start with workflows, business problems and measurable outcomes before deciding where AI fits.
Many organisations are starting their AI journey in the wrong place. They begin with tools. Should we buy Copilot? Should we use Claude? Should we build something custom? Should we standardise on one platform?
These are reasonable questions. They just should not be the first questions.
The first question should be: “What business problem are we trying to solve?”
Without that, AI becomes a technology purchase rather than a business change. And technology purchases, on their own, rarely transform organisations. A strong AI implementation strategy starts with the work, not the software.
Tools are visible. They are easy to demo. Easy to buy. Easy to announce. Easy to compare. They create the feeling that progress is happening. A business can roll out licences, host a training session, create a list of approved tools, and say it has started its AI journey. But tool access does not equal business value.
– If people do not know how to use the tools, value will be limited.
– If the tools are not connected to real workflows, value will be limited.
– If the data is not accessible, value will be limited.
– If no one owns adoption, value will be limited.
– If use cases are not tied to business outcomes, value will be limited.
The tool is only one part of the system.
The better starting point is the work itself.
– Where is the business slow?
– Where is it expensive?
– Where is it inconsistent?
– Where are people spending time on repetitive tasks?
– Where are decisions being made with incomplete information?
– Where is knowledge trapped inside individuals or teams?
– Where are customers waiting too long?
– Where are managers manually checking work that could be supported by AI?
– Where does growth require more headcount than it should?
These are better questions because they point towards AI business value. AI should be applied where there is friction, waste, risk, complexity, or opportunity.
That means the workflow comes first. The tool comes second. This is the difference between buying AI tools for business and building an AI adoption strategy that can actually change how work gets done.
One of the most practical lessons in AI adoption is this:
If you apply AI to a broken process, you may simply make the broken process faster.
This is why leaders need to understand the workflow before automating or augmenting it. That does not mean every organisation needs a six-month process mapping exercise before doing anything.
But it does mean teams need enough clarity to answer:
– What are the steps in the process?
– Who is involved?
– Where does information come from?
– Where are the delays?
– Where are the exceptions?
– What decisions are being made?
– Where does human judgment matter?
– What risk would increase if AI was introduced?
– What would improve if this worked?
Without that understanding, AI can create noise. A tool may speed up one task while leaving the wider process unchanged. It may create outputs that still require heavy manual review. It may be adopted by a few people but ignored by others. It may solve the visible problem while missing the real one.
Once the business problem is clear, tool selection becomes much more useful.
AI is not a single tool. It is a spectrum, and different problems require different approaches. You might solve daily bottlenecks like drafting, research, or summarising with a basic AI assistant. But scaling efficiency usually means embedding AI into existing tech stacks, like CRM, ERP, or HR systems, or securely linking models to proprietary data. Other times, the answer lies in workflow automation, custom development, or predictive models rather than generative text. Most importantly, some problems should not touch AI at all. Knowing when to walk away from the tech is a competitive advantage.
Sometimes, the highest-value solution has nothing to do with AI. A simple process change, a cleaner data structure, clearer ownership, or basic automation will often yield better, faster results. The strongest leaders today are not falling for the hype of trying to force AI into every single workflow. They are not asking, “How do we use AI everywhere?” They are asking, “Where is AI actually the right tool for the problem?”
There are common warning signs that an organisation is approaching AI as a tool rollout rather than a value creation effort. Look for:
– Lots of tools, but no clear use case ownership
– pilots that do not connect to business metrics
– employees using AI differently across teams with no shared standards
– low adoption after licences are purchased
– no clarity on what data can be used
– no measurement beyond anecdotal success stories
– no process for deciding which AI use cases scale
– leadership updates focused on tools rather than outcomes.
None of these mean the organisation is failing. They mean it is early.
But if they are not addressed, the organisation risks building AI activity without AI discipline.
Before buying or scaling AI tools, leaders should ask:
– What workflow are we improving?
– What is the current friction?
– Who owns the business outcome?
– What value are we targeting?
– What data does this depend on?
– What risk needs to be managed?
– Who will use the tool?
– How will they be trained?
– What metric proves this is working?
– What happens if the pilot succeeds?
– What happens if it does not?
These questions force clarity. They also help teams avoid buying tools for problems they have not defined.
A practical starting point is to create a simple AI opportunity map. For each potential use case, capture:
This does not need to be complicated. Its purpose is to turn AI from a set of ideas into a set of decisions. Once this exists, leaders can prioritise. They can decide which use cases are worth piloting, which require better data access, which need governance review, which require training, and which should wait.
That is the difference between tool adoption and operating discipline. It is also what turns an AI implementation strategy from a software decision into a business change plan.
AI is incredibly powerful and improving by the day, but tools do not create value. People do. True ROI happens when teams apply the right tools to the right problems, within the right workflows, equipped with the right skills, and guided by the right guardrails. The companies that dominate the AI era will not be the ones with the biggest software budgets. They will be the ones that slow down to ask better questions before they buy.
At Decoded, we partner with leaders and teams to build the judgement to do exactly that.
Build an AI implementation strategy that starts with the problem, not the tool. Take a look at our AI courses.
If you’d like to learn more about Decoded and how we can help transform your
organisation, we’d love to hear from you.
