We've not got our head in the sand here - we know that over the coming 3-5 years there will be some serious disruption to the job market via agentic AI. In fact, we're busy building out our own AI and process automation consultancy to deliver the outcomes for our clients that they will want to achieve with the rise of affordable AI tools.
To illustrate some of the things that are possible already, I wanted to write this blog to talk through how we're using AI at 1-Fix and to get you thinking about how this might apply to your business.
First thing to note - some housekeeping. If you don't already have an "Acceptable use of AI" policy, or similar, then you're off-track. Your team will be using AI tools (think Chat-GPT, Gemini, or the 'AI' buttons that are appearing in everything nowadays) without any guidance on what data (potentially commercially sensitive!) they are feeding into the systems.
It's always good to remember that if the product is free (a-la Chat-GPT non-paid plans) then you are the product - the data you're providing helps to train the models.
Don't be the trainer!
OK, now we've got the boring bit out of the way, here's a few of the ways we're using AI internally to lighten the load for our team.
Fixie - our AI ticket assistant
We've spent a number of months tinkering in the background to build out our AI powered helpdesk assistant that we've named Fixie (why not?!)
Fixie undertakes a whole bunch of tasks for the team, many of which are not customer facing so you won't have heard much from her, other than potentially a reply to you on a support ticket if you don't provide enough information for our techs to kick off the troubleshooting.
Here's some of the things Fixie does for us:
Looks at incoming support requests and puts them into the right service board (Helpdesk, Changes, Quotes, etc)
Determines the most applicable type/subtype/item for the service ticket (so we can report on, for example, how many printer issues we fix each month)
Automatically closes out certain types of tickets once it's run automated checks on the problem to see if it is fixed/can be auto-fixed, saving technician wasted time
Reads the incoming support tickets and determines if there's enough information for a technician to start work on the problem. If not, it responds to the client asking them for more info on the issue
Provides a summary of the ticket if it's escalated to a senior, so they don't have to read pages of notes from other techs - it's all summarised neatly for them with suggested next steps
Spots long running tickets and potential projects and reports them to our service delivery manager
We've got lots more to add to Fixie over the coming months, but these are all "quality of life" improvements to take away repeat and potentially wasted tasks from our team.
Fixie is back-ended by a service called N8N, which is a powerful workflow builder. We've used workflows into our ticketing system (ConnectWise PSA) and our monitoring system (NinjaOne) along with some AI calls to OpenAI to make Fixie. Fixie conservatively saves our team around 15 hours per week in manual tasks at present,
Replit - Rapid application development
Many people won't know that my background in IT actually started in programming. I used to develop systems in Basic (and Visual Basic) and could also hack together a bit of C++ (badly). However, the rise of AI models that can write pretty decent code has lead to something quite incredible - something called Vibe Coding.
Vibe Coding is building websites, web apps and even mobile apps by simply prompting a chat bot in the same manner you would prompt Chat-GPT or Gemini. Of course, there is still a very big dollop of "Garbage in, garbage out" in the sense that bad prompting will either give you an unexpected result, or more often than not an application that is full of security holes or problems you end up spending more time (and money) to try and fix.
However, when it works, it's something quite magical.
Last month we unfortunately said goodbye to Bonnie, our Service Delivery Manager. One of Bonnie's tasks was to allocate the tickets to the technicians. This is time-critical, and with 1-Fix receiving anything from 20-60 tickets per day, it means that it's hard to be present in meetings, take a break, or generally relax due to the need to be ready to pass a ticket to the appropriate engineer at the right time.
Due to the challenges of covering this while we await her replacement starting, I turned to my vibe coding app of choice - Replit - and spent a Sunday afternoon building an automated system that can allocate the tickets to the tech team.
The system has all sorts of 'smarts' - it can understand engineer ticket load, skills, and calendar entries and will allocate the work based on who is best placed to deal with that individual ticket.
It's taken around 6-8 hours of time prompting and refining the software, but v1 was up and running on the Monday following the initial build and it's saved our sanity.
We won't use it to replace a job role - but we will keep the tool (and keep refining it) to allow the new Service Delivery Manager to focus on training, service improvement, client relationships, and team dynamics... all the good stuff that humans do well.
Coming Soon....
I've recognised the need for 1-Fix to be able to provide these AI and smart process automation services to our clients, but I also know that I'm not the best person to be doing it - it's not scalable or efficient.
In order to tackle this, we're currently working hard to build out our own automation arm of the business to help solve our clients challenges and help them embrace the power and efficiency improvements that AI and automations can bring to their businesses.
We already have a wait-list of interested clients who are keen to know how to start on their AI journey. If you'd like to be part of the story, register your interest here and I'll update you once we're ready to start working with our first batch of customers in Q2 2026.