Skip to main content
  1. Services/

Generative AI

746 words·4 mins
AI Operations
Table of Contents

Generative AI (“GenAI”) is a term used to describe the most recent wave of AI models which are able to generate content from a relatively short prompt, which is usually text based, or increasingly multi-modal (using images, audio or even video).

Why should I care about AI?
#

At it’s simplest level, generative AI opens up opportunities to introduce deeper automation into tasks that have traditionally been performed by knowledge workers in the same way that mechanisation introduced automation into manual labour during the industrial revolution. We haven’t totally removed human work from industrial production, but those roles have changed with the increase in mechanisation. In the same way, we won’t totally eliminate human work in knowledge-based roles, but those roles will change.

As of 2025, many organisations are already finding significant productivity gains from embedding modern AI into their workflows, especially when it comes to software development, content generation and planning.

But my business isn’t really in “tech”
#

It doesn’t matter. The biggest paradigm shift in the last couple of years has been easy access to AI models which are generalist, which means they have general knowledge about a range of subjects. As a very easy example, take a new member of the team who needs to write a presentation, AI can support them in structuring their work so they can communicate better. It may even be able to help generate relevant imagery and diagrams for them faster. Whether your business is in “tech” or not, AI is applicable to every role and can be used as a productivity multiplier on your team.

I’m in a small business, how deep do I need to go on this?
#

If you’re a small business just getting into AI, there are a few easy things you definitely don’t need to do (and you should be skeptical of any solutions provider who says you need to):

  • You don’t need to build an LLM. Especially, you don’t need to build what’s called a Foundation Model. The up front cost to build one from scratch, to a quality that is generally competitive, is in the tens of millions of dollars. There’s also a strong argument, that the first mover advantage of companies like OpenAI, Google, Anthropic and Meta; mean that new entrants may never be able to compete without very very deep pockets.

  • You don’t need to hire a full time, permanent, “AI Expert”. Any of the current experts, are working with a technology which is evolving faster than their sales pipeline. Last year’s “expert” is already obsolete.

What you should do, is start getting AI tools into your organisation. Some may already be available to you (for example Google Gemini is available as part of many Google Workspace plans). Start your team developing the skills to work with AI tools and get confident with them.

There is a quite clearly defined skill in “telling the AI what you want”, especially in a way that gives you helpful answers. More generally you could call this “Prompt Writing”, “Prompt Engineering” or just “Prompting”. It very closely models how you might give instructions to an outsourced agency contractor, or new grad hire to your team. It’s a fairly safe bet that this is already THE fundamental skill that all others are built on when it comes to effectively working with AI, and that the sooner you start developing this across your entire team, the more options you will have as the landscape evolves.

How can I actually generate savings through AI?
#

It’s important to remember that AI is a productivity tool, in that it can increase the output of your current tooling and team. As with any productivity gain, it can only lead to savings if you use those gains to spend less. Especially true when you account for the increased operational costs which come from paying for the AI tools themselves.

One easy way to get started, is then to consider putting steps in your hiring and procurement signoff processes to ensure no headcount or tooling spend is approved unless the team has aleady demonstrated they can’t fill the need through AI productivity gains. Only by spending less money in other places can you expect to see an overall cost base improvement from an increased focus on AI.

Want to talk more?
#

If you want a pragmatic conversation about how to use GenAI in your organisation, and how it may change the roles of your team. Let’s chat.

[Alan Cruickshank](/about/alan)
Author
Director @ A14K | Author & Maintainer @ SQLFluff | DataIQ 100 2020.