AI & Automation

#

Knowing the difference will save you time & money.

#

AI & Automation are powerful tools that can benefit any business of any size but they can also be a money-pits if you don’t have a clear understanding of the options and the ROI you would like to achieve.

Knowing where to start is never easy but made so much harder by the outlandish claims coming from the big AI industry players as they race for dominance and ever more investment.

Big tech likes to throw around big questions like:

• Will it take or enhance your job ?
• Will it create new opportunities ?
• Will it destroy or revolutionise your business ?

But conveniently skip the crucial ones:

• What can AI actually do ?
• Will it give you ROI ?

They do this because they don’t really know themselves. The dream being sold is just one aspect of AI that can act independently like a human when the real ROI for most of us lies in much simpler automations and ai-automations. Think of a washing machine, it’s not clever but it is so much better at washing clothes than us humans. Why buy a robot that may or may not be able to wash clothes, do the dishes and clean when a washing machine, dish washer & hoover will do that in a more reliable & cost-effective way ?

Automations

#

Automations are simple bots, they don’t use LLMs & don’t cost much to build or run. They’ve become the overlooked child of the industry because they already exist and already work so can’t attract big investor money.

Like a washing machine they’ll do one thing far better & quicker than a human. They aren’t going to change life as we know it but they’re a reliable tools that are cheap to run and will certainly produce ROI.

An example of a typical payroll automation is a bot that downloads payslips from a payroll provider then uploads those payslips to an employees account saving time and reducing human error.

AI Automations

#

AI Automations also do one thing but make use of ai tools to get some or all of the task done. They cost more to run as they make use of AI tools like OCR and LLMs to scan images or extract useful data. They also do one thing well so reliably and consistently produce ROI.

An example of a typical accounts automation would be a bot that downloads invoices from many different suppliers, uses AI tools like Optical Character Recognition and LLMs to extract the core data from the invoice before putting that data into a spreadsheet or maybe sendiing it straight into an accounts package.

AI Agents

#

These are more complex and the dream we all hear about on Social Media and linkedin. Rather than bots they’re refered-to as agents because they use tools like LLMs to make decisions and accomplish multiple tasks often working in groups together. That sounds complex because they are. They’re also costly to build and the source of comically-gross errors we all read about. ROI is hard to call and the statistic you’ve probably read that 95% of corporate AI initiatives fail - that’s mostly down to AI agents.