There's a version of the AI conversation that's become background noise by now. The one about automation, displacement, and whether machines are coming for jobs. It's not a useless conversation, but it's also not the one that matters most to people running a 20-person consultancy or a 40-person agency right now.
The more immediate and practical question is this: is your business structured in a way that can actually get value from AI? Because the answer to that question has almost nothing to do with AI, and almost everything to do with your data.
Every AI assistant, whether it's built into your HR platform, your project tool, or a standalone chat interface, is fundamentally constrained by the data it has access to. Ask it a question about a domain it can see, and it performs remarkably. Ask it a question that requires data from a domain it can't see, and it either guesses or fails.
This is why the most useful question to ask before evaluating any AI feature isn't 'how smart is the model?' It's 'what data does it have access to?'
Consider a question that most agency leaders want to answer on a regular basis: 'Who should work on this project?'
To answer that intelligently requires:
An AI assistant that can only see your HR data can tell you who's on leave. One that can only see your project tool can tell you what's already assigned. Neither can answer the full question. Only a system where all of that data lives together, and an AI layer that spans all of it, can give you a genuinely useful answer.
Right now, every software vendor is adding an 'AI' badge to their product. Most of it is genuine; AI is genuinely improving summarisation, drafting, and workflow automation across the board. But there's a pattern worth being sceptical about.
When AI is bolted onto a siloed tool, it can only answer shallow questions. It can summarise what's in that tool. It can draft content based on what's in that tool. It can generate reports from what's in that tool. All of that is useful at the margins.
But the questions that actually matter in a knowledge-based business aren't shallow. They span domains:
Shallow AI, operating on siloed data, cannot answer any of these questions. Deep AI, operating on connected data, can answer all of them. The difference isn't the sophistication of the model. It's the breadth of the data.
The businesses that will get the most value from AI over the next few years are not necessarily the ones investing most heavily in AI tooling. They're the ones that have done the less glamorous work of getting their operational data into a single, coherent place.
Think of it like this. AI is a very powerful engine. But an engine needs fuel to run, and the fuel here is data. High-quality, connected, real-time data about your people, your projects, and the time your team spends on both. Without that, even the most advanced AI model produces noise.
For agencies and consultancies specifically, being AI-ready means three things:
Businesses that have those three things in place will find AI genuinely transformative. Businesses that don't will find themselves paying for AI features that can only answer the easy questions — and those were never the ones costing them.
If you're evaluating AI features in any platform, ask one question before anything else: 'What data does this AI have access to?'
If the answer is 'just what's in this tool', you're looking at shallow AI. If the answer is 'all of your people, project, and time data simultaneously', you're looking at something that might actually change how your business operates.
The gap between those two things is not small. And over the next few years, it will become one of the more significant operational advantages available to knowledge-based businesses.
What's one question you wish you could ask your business, about your people, your projects, or your performance, and get an instant, accurate answer to? That question is a useful diagnostic for whether your current data setup is working.
Workup connects your people, projects and time in one intelligent platform.