From dashboards, to conversations
TAG Catering Equipment spent years building the bespoke reporting it actually needed with Tugger®. Now Dean Kahl asks Claude the questions those dashboards can’t answer, straight from TAG’s Simpro® data. And he reckons he might never need another bespoke report built.
Three systems that didn’t talk
We’ve told part one of this story before. Worth a quick recap, because the direction of travel is the point.
TAG Catering Equipment designs, supplies and maintains commercial kitchens - everything from local restaurants to five-star hotels and Michelin-starred kitchens. Around 50 people, based in Stevenage, 30-plus years in the trade.
When Dean Kahl joined in 2019 as marketing manager, the business was running on three systems that didn’t talk to each other.
“We had our quoting platform, an industry-wide platform called AutoQuotes, primarily for projects and sales. We had our job management system - an old, antiquated, server-based system for our service and maintenance division. And then we had Sage.”
“I come in as a marketer and asked the question ‘how do I get a one customer view, or understand what all our customers are doing across both divisions of the business?’”
Covid forced the issue. TAG moved to Simpro. And then hit the wall that brings most people to us.
“We quite quickly realised that the reports within Simpro weren’t giving us the information we wanted, or representing the information how we needed it represented.”
Reports that think like TAG
Their Simpro account manager pointed them towards Tugger. The generic Power BI reports got them close, but not all the way, so Dean came to Craig and asked for bespoke.
We've invested a significant amount in bespoke reports over the years. They continue to be used and offer huge value to our business insight.
There’s a sales report, a service desk report, a marketing report. The point was never to have reports. It was to have reports that matched how TAG thinks.
“The idea of those reports was to create reports that would be useful for us and the things we wanted to report on - as opposed to, this is how Simpro displays it, but that’s not quite how we want to report on it.”
The clearest example is First Time Fix. It matters enormously to a service and maintenance business, and it’s a metric every company defines slightly differently. That’s why it doesn’t come ready-made out of any job management system.
“Our concept of First Time Fix is very different to the next company’s concept of First Time Fix,” Dean says.
So TAG defined their own, and Tugger’s job was to make the data support it. Dean and Gabor have been through a few revisions to get there. That’s the real work - not pulling the data, but agreeing what the number actually means in your business.
“What Tugger has enabled us to do is join the dots, join the pieces of the puzzle of the Simpro database, to give us an example of what First Time Fix is.”
The questions a dashboard can’t answer
TAG had solid, scheduled reporting the team used every month. So what pushed them towards AI?
The questions people started asking.
“I started getting asked rather specific questions, conversational questions of data. Whether that’s born from people using AI, or Google, or just our business requirements changing.”
Some of those questions weren’t worth a bespoke report. You need the answer once.
“AI Insights and the MCP into Tugger has enabled us to query the data that isn’t necessarily in existing BI reports. Those things you only need to query once, or those nuanced queries that you can’t really get out of Power BI. As much as we’ve added lots of filters to our BI reports, sometimes you still can’t get to the end result you need.”
That’s the gap.
“The Claude integration with Tugger accelerated the ability to get those custom pieces of information out.”
And here’s Dean on where this is heading. It’s the quote we probably shouldn’t like, but do.
We might never come to Tugger for a bespoke Power BI report ever again. Because AI has the ability to do the job.
We’ll take that. Building bespoke reports one at a time was never the endgame. Giving people the data structure and the tools to answer their own questions is.
One question, two routes, the same number
Dean raised the thing everyone raises about AI and data.
“We spent a lot of time developing the Power BI reports, interrogating the data, making sure it’s accurate. Then we know it’s accurate at scale. But there’s always that question of - is it accurate?”
Fair question. So TAG tested it properly. Here’s how.
The question. TAG’s Power BI numbers are proven. Would Claude match them?
The prompt. Gabor at Tugger writes a prompt covering exactly how TAG defines First Time Fix.
The test. Dean runs the same date range through Claude and through the report TAG already trusts.
The result. The same number, by two different routes.
He said, right, put this into Claude with the same date range, and see if it comes to the same result. If it does, you know you're on the right path.
Worth being clear why this works. Tugger isn’t asking the AI to do the maths. The calculations happen on our side against your structured data, so the AI is served an accurate answer.
Your data already knows. AI Insights just reports it back.
The definitions, not the connection
TAG haven’t rolled AI querying out across the business yet, and Dean’s reasons are thoughtful ones.
Some of it is access. Financial data, margin, profit - only certain people see that in Simpro, and any wider rollout has to respect the same boundaries.
Most of it is training.
It's moving at such a fast pace, and you don't want people equipped with information that's inaccurate and then speaking to customers about it. There are certain parameters they perhaps should be asking about - cost centres, job type - that they're not aware of, because they think AI is psychic.
First Time Fix makes the point again. If someone just asks “what’s the First Time Fix on this?”, the answer needs to come back the same way every time.
“You can set that up in a project, so if anybody ever queries it, it’s looking at the same parameters.”
Which is the real work of getting an organisation AI-ready. Not the connection. The definitions.
“It will come. It will undoubtedly come.”
What other businesses can take from this
TAG’s path is the one we think most Simpro users will end up walking.
Get your data structured. Build the reports you actually need. Then let AI handle the questions the dashboards were never going to answer.
Dean’s already there. Next on his list, scheduled checks and recurring reports that run themselves. And his ask of us isn’t another dashboard.
“I haven’t got the time to set up the project. I haven’t got time to set up the queries. Could you now support us with setting up accurate prompts and projects for our reports to run?”
That’s the shift for us too - from building the thing to helping people build their own. Less report writing, more prompt and project design.
You don’t have to do it in that order, and you don’t have to start big. But you do have to start.
