Clektive Thinking: A practical guide to implementing AI in your business

AI is evolving at an incredible pace. New models, tools and agents are appearing all the time, and businesses are under increasing pressure to work out what all of this could mean for them. 

For CEOs, COOs, CFOs and CMOs, the challenge is not necessarily understanding every new development. It is working out where AI could create genuine value for your business, deciding what approach to take, and turning the opportunity into something practical that works. 

In the latest episode of Clektive Thinking, Chris Harling and Clekt’s Solution Architect, Elliott Fairhall, explore a practical framework for thinking about AI in business, from identifying the right opportunities through to choosing the right technology and understanding the different types of AI now available. 

In this guide, we explore: 

• Where AI can create value across smarter employees, faster processes and transformational products 
• What to do once you have identified an opportunity, including whether to use an existing tool, buy a solution or build something yourself 
• What predictive, generative and agentic AI actually mean and how they can be applied in practice 
• How to bring these ideas together to identify practical AI opportunities 
• Why starting with the problem, rather than the technology, is so important 

1. Start with the problem, not the technology

It is easy to start an AI conversation with the technology. 

Which model should we use? Should we be using Claude? What about ChatGPT? Do we need an AI agent? Should we build something ourselves? But that risks getting ahead of the most important question: what are we actually trying to achieve? 

The first step is to identify a problem or opportunity in the business. That might be something that is taking too much time, a process that is unnecessarily slow, information that is difficult for people to access, or an opportunity to create something that was not previously possible. 

Once you understand the problem, you can start to think about where AI could help. 

As Elliott explains: “What do you want it to do for you? Not necessarily, I just want the AI.” 

2. Think about where AI could create value

Elliott describes three broad areas where businesses can look for opportunities: 

1. Smarter employees 

AI can give people better access to information and help them do their jobs more effectively. This could mean helping employees find and analyse information more quickly, supporting knowledge workers with research and writing, or giving teams access to information that would previously have required support from another department. 

2. Faster processes 

AI can also be used to improve processes that are slow, repetitive or cumbersome. The opportunity here is not necessarily to replace an entire process. It could be about removing some of the manual work, speeding up individual stages or helping people make decisions more quickly. 

3. Transformational products 

The third area is about using AI to create something new. That could mean a new customer experience, a new product, a new service or an entirely new revenue stream that would not have been possible without the technology. 

These are not three boxes that every business needs to work through in order. A single opportunity could involve more than one of them, with people, processes and technology all playing a role. The important thing is to understand the problem first, then work out where the opportunity sits and how success will be measured. 

3. Ask whether AI is actually the right solution

Identifying an opportunity does not automatically mean you should solve it with AI. 

Sometimes the problem is that an existing system is not being used properly. Sometimes a process simply needs to change. Sometimes a more straightforward piece of technology will do the job. That is why the next decision is about choosing the right approach, rather than automatically choosing AI. 

Elliott makes this point clearly when talking about the range of tools now available: 

“So, it’s choosing the right tool for the right job then.” 

That could mean using an existing AI application, introducing a tool such as Claude or ChatGPT, using a developer tool such as Claude Code, or building AI directly into an application or business process. 

The decision to build something yourself should therefore come after understanding the problem and the available options, not before. 

There is also a danger in assuming that because something involves AI, it must automatically be better. 

A process that works well today may not need to be replaced simply because a new technology has become available. Understanding the current process, measuring how well it works and defining what improvement would actually look like gives you a much better basis for deciding what to change. 

4. Understand the different types of AI

Once you have identified an opportunity where AI could genuinely help, it is useful to understand what type of AI might be appropriate. 

Three terms come up regularly: predictive, generative and agentic AI. They describe different capabilities, rather than three stages that every business needs to move through. 

1. Predictive AI 

Predictive AI uses historical data to predict what might happen next. Common examples include demand forecasting, fraud detection, churn prediction and recommendations. 

For predictive AI to work well, the underlying data matters enormously. It needs to be clean, organised, accurate and representative, with enough historical information to identify meaningful patterns. If the data is poor, the prediction will be poor too. 

2. Generative AI 

Generative AI creates new content such as text, images or code. For businesses, this can be particularly useful for knowledge work. It can help people analyse information, create first drafts, summarise material or produce content much more quickly. 

But there are limitations. Generative AI can produce inaccurate information, lose a company’s tone of voice or create content that still needs human review. The opportunity is therefore not simply to hand a task over to AI, but to understand where it can make people more productive while keeping appropriate controls around the output. 

3. Agentic AI 

Agentic AI goes a step further. Rather than simply providing an answer, an AI agent can work towards a goal, making plans, reasoning through tasks, using tools and interacting with different systems. 

That creates significant opportunities for automation, but it also introduces more risk and uncertainty. The more autonomy an AI system has, the more important clear boundaries and guardrails become. 

5. Bring the frameworks together

One of the useful points from the conversation is that the three value areas and the three types of AI are not directly linked. 

Smarter employees, faster processes and transformational products help you think about where AI could create value. 

Predictive, generative and agentic AI help you understand what AI can actually do. 

The two can then be combined. 

For example, a business might identify a research process that takes employees several hours every week. Generative AI could help with the initial research and analysis. Predictive AI could process structured information and identify patterns. Generative AI could then turn those findings into a report or presentation. An agent could potentially connect those steps and manage the workflow. 

The technology can become quite sophisticated, but the starting point remains the same: understand what you are trying to achieve. You do not need to tackle everything at once. In fact, trying to do so can make it much harder to demonstrate whether your investment is actually delivering value. 

6. You do not need to automate everything

There is a tendency in conversations about AI to jump straight to automation. But not every opportunity needs to be fully automated. Sometimes the biggest benefit comes from helping people do their jobs better. Sometimes it is about making information easier to access. Sometimes it is about removing a small but frustrating part of a process. 

The important question is what will make a meaningful difference to the business. That also means thinking about measurement from the beginning. If you cannot explain what success looks like, it becomes difficult to know whether an AI project is actually working. 

Rather than starting with “where can we use AI?”, a more useful question might be “where are we spending time, money or effort that could be improved?” 

7. Put the foundations and guardrails in place 

The more advanced the AI application, the more important the foundations become. 

Predictive AI relies on good quality data. Generative AI needs appropriate controls and human oversight. Agentic AI needs clear boundaries around what it can access and what actions it is allowed to take. That means AI implementation is not simply a technology project. 

It can involve data quality, processes, governance, people, technology and measurement. Getting those foundations right helps businesses move from experimentation towards something that can actually be used at scale. 

The takeaway

The conversation around AI has moved quickly from understanding what the technology can do to working out how businesses can actually put it to work. 

As Chris puts it towards the end of the conversation: 

“It’s kind of gone from AI insight to AI action.” 

And that shift makes the starting point even more important. For Elliott, the question businesses should be asking is not which AI tool they should buy or which model they should use. It is much more fundamental: 

“What problems are we trying to solve? That’s what matters right now.” 

Once you understand that problem, you can start asking what AI could do to help solve it. The technology will continue to change. New models, tools and capabilities will keep appearing. 

The business problem, however, is where the conversation should begin. 

Ready to put AI to work?

Knowing where AI could make a difference is one thing. Turning that opportunity into something practical is another. 

If you’re exploring how AI could help your people, improve your processes or create new opportunities for your business, Clekt can help you work out where to start and what to do next.