AI Makes New Things Viable, Not Just Existing Things Faster
AI makes new things viable, not just existing things faster. The challenge is spotting which opportunities are worth pursuing.
“We could automate that.” It’s a reasonable instinct. Process takes too long. Tool makes it faster. On to the next one.
But that instinct has a blind spot. It assumes the work you’re already doing is the right work. Just too slowly. The real advantage goes to the businesses that pause long enough to ask a different kind of question, and then notice opportunities that were always technically there but never quite worth pursuing. The time, cost or complexity involved made them impractical. AI changes that calculation. Not by making you faster at what you already do, but by shifting the economics of what you can offer.
Services that couldn’t turn a profit. Analysis nobody had the weeks to do properly. Ways of working with customers that would have needed a team you didn’t have.
The difference between faster and new
Making something faster is valuable, but it’s still the same thing done with less friction. Making something newly possible is where the real AI opportunities sit.
Faster
Doing what you already do, more efficiently. Generating reports quicker. Automating admin. Speeding up research. Real gains, but the work itself hasn’t changed.
New
Doing things that weren’t practical before. Offering personalised insight at scale. Analysing years of customer data to spot patterns. Creating services that only work because the economics have shifted.
The right side is where strategic advantage lives, but it asks you to think about your business differently, not just run it with less effort. That’s why it gets walked past.
Efficiency questions versus judgment questions
The difference shows up clearly in practice.
Efficiency questions
How do we do this faster? Can we speed up our reporting? Can we generate proposals quicker? Can we reduce the time spent on admin?
Judgment questions
What should we be doing differently? What do our customers need that we’re not offering? What patterns are we missing in our data? What would we build if speed wasn’t the constraint?
If your AI conversations stay in the left column, you end up optimising a business model that might already be under pressure rather than evolving it. The right column is harder to get to because it requires stepping back from the day-to-day. That’s not a prompt engineering problem. It’s a judgment call.
The most valuable thing AI can do for your business might not be answering your current questions faster. It might be revealing that you’re asking the wrong ones.
Where to look
These opportunities don’t announce themselves. They sit in the space between what you currently offer and what your customers actually need.
What your customers haven’t asked for yet. Every business has a list of things they know customers would value but could never offer at a margin that worked. Bespoke analysis, ongoing monitoring, highly personalised recommendations. AI changes that maths, making it realistic to deliver things that would have previously needed a dedicated team. What problems sit adjacent to the work you already do? If you could offer something new tomorrow, what would it be, and why don’t you offer it already?
Data you have been sitting on. What would you want to know if you could ask anything of your own data? Years of project histories, customer feedback, pricing decisions, competitive intelligence. Most of it buried in spreadsheets, inboxes and shared drives because nobody had the weeks to make sense of it. Now you can surface patterns in hours, and those patterns can become the basis for conversations with customers that you simply could not have had before.
Gaps between your services. After you deliver a piece of work, before they come back for the next one, there is usually a silence where your customer is on their own. AI lets you fill that silence with something useful: ongoing insight, monitoring, lightweight guidance. It turns a project-based relationship into something continuous.
Your business model itself. Where does your current model create friction for the people you’re trying to serve? What would you change about how you price, package or deliver your work if you weren’t constrained by how things have always been done? AI doesn’t just make existing services faster. It can make entirely different service shapes viable for the first time.
Sustainability check
Some of the most valuable opportunities sit at the intersection of AI and sustainability. Supply chain visibility that wasn’t practical before. Energy monitoring that used to need a dedicated team. Circular design informed by real usage data. When you start asking “what should we be building?” instead of “how do we speed this up?”, you also start asking “what’s the real cost of this?” The best AI questions have sustainability built in, not bolted on.
How to tell signal from noise
Not everything that’s become viable is worth doing. AI makes a lot of things possible that still aren’t good ideas. Just because you can build it doesn’t mean someone needs it. The best test is whether your customers would pay for it without you having to explain why it’s clever.
A few filters that help:
Does it solve a problem your customers already have? The best AI-enabled services aren’t solutions looking for problems. They’re answers to questions your customers have been asking for years, ones you couldn’t previously address at the right price or speed.
Would it change how customers see you? If it reinforces what people already think, it’s probably an efficiency play wearing a different hat. The interesting opportunities shift perception.
Can you deliver it without losing what makes you good? The strongest opportunities build on what you’re already known for rather than asking you to start again. Be honest about whether you’re extending a strength or abandoning one.
Why the right questions don’t get asked
The biggest barrier isn’t technical. It’s cultural. Most organisations reward people for improving what exists, not for imagining what doesn’t yet. Suggesting a genuinely new service feels risky in a way that suggesting a faster process never does. So people don’t.
There’s pressure to show quick wins, to demonstrate ROI, to have a clear use case before you start. Efficiency-first thinking gives you all of that. The harder questions, the ones about what your business should be building or offering or changing, don’t come with tidy metrics attached. They require sitting with uncertainty, and most organisations aren’t set up for that.
The quality of what you build depends on the quality of what you ask.
If you want your team to spot these opportunities, they need permission to bring half-formed ideas to the table. The best thinking at this stage won’t arrive as a polished business case. It’ll arrive as a hunch, something someone noticed while doing their day job. The best ideas often start as a throwaway thought in a ten-minute chat, not a strategy document.
A way to start
Pick one customer relationship, one data set, or one part of your business model and spend thirty minutes with this question: if AI removed the constraints we’ve always worked around, what would we do differently?
Don’t try to answer it fully. Just notice what comes up. The interesting answers usually aren’t the first ones. They’re the ones that arrive after you’ve got past the obvious efficiency plays and started thinking about what your business could become rather than what it currently is.
Then ask a harder question: what would your customers actually want this to be? What they need and what you’ve been delivering may have quietly moved apart, and AI gives you a way to close that distance that simply wasn’t there before.
What’s next
Seeing what’s possible is one thing. Knowing when to let AI run with it, when to keep a human hand on the wheel, and when to leave it alone entirely is something else. That’s the judgment we’ll explore in Part 3: When to Delegate, When to Lead.