Our customers often ask us about AI and whether they should be using it. There are certainly lots of use cases for it, we have even written AI powered tools for ourselves and our customers.

Here is how we currently use AI;

Use AI to change data formats

We are often provided information in a certain format, however we more often than not need it in a different format. AI is awesome for this. Simply provide the data and use commands such as

• the provided list contains values and IDs, place the IDs first, followed by the values separated by a comma. Each item should ne on a new line
• the following data features full names and dates in dd/mm/yyyy format. Split the names into firstname and lastname and change the date format to mm/dd/yyyy

Previously we’d either use Excel or write a script, but now we can just ask AI to modify the data into whichever format we need.

AI in our own software

Many of our clients are in the recruitment sector. Recruitment is heavy on data and as you saw above, AI is great with data.

We have created AI scripts to make life for recruiters easier, including

• We provide a job spec, we provide info of all candidates who have applied, AI then sorts the candidates into order of who fits the role the best, explaining its reason for each candidate position
• Create a summary document for recruitment agencies to provide to their clients. The summary will contain high level info such as the number of applicants, but also drill down to the quality of applicants and make further suggestions
• Create job packs for each job, providing an overview of the job in general, the skills required and candidate expectations
• Similar to assessing applicants, AI can also look through recruiters entire candidate database in seconds to headhunt the perfect candidate
• Make suggestions on human written content before publishing. For example, if the position is for a ‘web designer’, did you remember to include the web design software you need the designer to be competent using?

AI can get things wrong however, we always write additional checks into AI powered software. The last thing you want is for your content to be destroyed by an over eager AI.

Initial Code Debugging

If an error is produced, sometimes it can be a needle in a haystack trying to find it. Website owners can add a multitude of plugins, all running their own code. If we don’t initially recognise an error, AI can sometimes help.

“website domain.com is producing the error ‘xx at yy’”. AI sometimes can be a great help and respond with “This is a common issue with ABC Plugin after their latest update. I see this plugin is running on this website. Try disabling the plugin and see if the error persists. If not, it’s certainly the plugin causing it.”

AI debugging won’t be much use helping fix the error, but can speed up identifying the problem.

What about writing code?

Not a chance, of course we have played with AI’s code writing ability but it is no where near good enough to replace a developer.

Our code is clean and efficient, this results in a fast loading website.

AI often produces heavily bloated code littered with bugs. It’s like building a race car with tape. It might kind of work for a short while but all hell is about to break loose.

There are lots of AI website creators these days. Right click on the page and select View Source. You’ll see a mass of gobbledegook most likely, that’s also what search engines see.

Now View Source on one of our websites. You will see your page content with significantly less extra code around it.

We had several AI tools create one of our website designs, and compared the output to our developer code. Our developer used 1294 lines of code in the main template, with the entire template using 42kb.

The AI website used tens of thousands of rows and weighed in at over 8MB. That’s a 19400% increase!

The AI site also had many issues that needed sorting, the resulting website did not match the design properly which would require more code to fix.

Summary

We use AI for various benefits and it works extremely well. AI is great for data tools that would take a human many hours to achieve.

But be wary, as AI can and often does get things wrong, so have a backup plan to make sure the output is good before using it.