Every business owner I speak to is currently staring at the same blinking cursor. You’re being bombarded by software updates, LinkedIn thought-leaders, and sales emails all claiming to have the 'magic bullet' for your operations. The pressure to innovate is real, but it’s often masked by a much louder question: should I use AI in my business right now, or is this just another cycle of expensive software bloat?
I’ve watched thousands of businesses navigate this transition. I’ve seen companies save £50k a year by switching to a single AI-native tool, and I’ve seen others burn through six-figure budgets on legacy software that simply added a 'Generate with AI' button to a broken process. The difference between these two outcomes isn't luck; it’s the ability to distinguish between AI-Native tools and what I call SaaS Taxidermy—legacy software that’s been stuffed with an AI API to make it look like it’s still alive in the modern era.
To help you cut through the noise, I’ve developed a framework. Before you sign another seat-based contract or upgrade your tech stack, ask these five questions.
1. Does it collapse the workflow or just change the interface?
💡 Want Penny to analyse your business? She maps which roles AI can replace and builds a phased plan. Start your free trial →
This is the most critical distinction when deciding should I use AI in my business for a specific function.
Legacy 'AI-wrapped' tools usually improve the interface. They give you a chatbot on the side of the screen that helps you write a description or summarise a thread. That’s helpful, but it’s incremental. It saves you three minutes of typing but keeps you trapped in the same manual workflow.
AI-Native tools collapse the workflow.
Instead of a tool that helps you write a better invoice (interface improvement), an AI-native tool monitors your project management software, detects when a milestone is hit, generates the invoice, reconciles it against your bank feed, and only pings you if there’s a discrepancy. The workflow hasn't been improved; it has been deleted.
When evaluating a new tool, ask: Does this remove steps from my day, or does it just make the existing steps slightly faster? If it doesn’t delete a task, you’re likely paying for a wrapper.
2. Is it 'Baked In' or 'Bolted On'?
I see this pattern across every industry, from retail to professional services. Legacy SaaS companies are desperate to avoid obsolescence. Their solution is often to 'bolt on' a third-party AI model (like GPT-4) as a separate feature.
This creates what I call The Latency Gap. Because the AI is an afterthought, it doesn’t have deep access to the core data of the software. It can only 'see' what you copy-paste into it or what it pulls through a limited API.
An AI-native tool is built around the model. The data structure, the user experience, and the automation triggers are all designed with the assumption that an LLM is doing the heavy lifting.
For example, if you’re looking at optimising your software costs, look for tools where the AI is the engine, not the paint job. A bolted-on AI feature is usually a sign that the company is trying to justify a price hike rather than fundamentally evolving their product.
3. What is the 'Unit of Value'?
For twenty years, SaaS has been sold by the 'seat'. You pay £20/month for every human who logs in. This model is fundamentally at odds with the AI era.
If a tool is truly AI-native, it should be reducing the number of humans needed to log in. If a software company is still pushing a heavy seat-based model while claiming to be 'AI-first', they are betting against their own product's efficacy.
When you ask yourself, "should I use AI in my business?", you should also be asking: "How am I being charged?"
I’ve found that the most effective AI tools are moving toward Outcome-Based Pricing. You pay for the task completed, the invoice processed, or the lead qualified. This aligns the software’s success with your own. If you’re still paying for 50 seats for a tool that claims its AI 'does the work of 10 people', the math doesn't add up.
4. Does it learn from your context, or is it just 'General Purpose'?
Many 'AI' tools are just wrappers for general-purpose models. If you ask a general AI tool to write a marketing plan for your specific plumbing business in Leeds, it will give you a generic answer.
Real AI value comes from Contextual Intelligence. An AI-native tool should be able to ingest your historical data, your brand voice, your pricing structures, and your customer feedback to provide answers that are unique to you.
This is why I often steer businesses away from generic 'AI for everything' platforms and toward specialised tools that solve one problem deeply. Whether it's IT support automation or legal document review, the tool must prove it can handle your specific edge cases, not just the 'average' case.
5. The 90/10 Rule: What happens to the remaining 10%?
This is where most business owners get stuck. They see an AI tool that can handle 90% of a job—say, bookkeeping or basic customer service—and they hesitate because of the remaining 10% that requires human judgment.
My advice? Automate the 90% and treat the 10% as a management task.
The mistake is keeping a full-time human (or a manual legacy system) to handle 100% of the work just because the AI can't do the last 10%. When you adopt an AI-native tool, your role shifts from 'Doer' to 'Editor'.
You shouldn't be looking for a tool that is 100% perfect. You should be looking for a tool that handles the bulk of the volume so your humans can focus on the high-value, complex edge cases. If you're still managing your business through manual tracking, consider how AI-first approaches compare to traditional spreadsheets in terms of error rates and speed.
The 'Agency Tax' and Why It Matters Now
One of the biggest 'non-obvious' observations I’ve made lately is the emergence of the Agency Tax. Many businesses pay agencies £3,000/month for execution work (writing SEO posts, managing ads, basic reporting) that can now be handled by £50/month AI-native tools.
The agencies aren't necessarily lying to you—many of them simply haven't updated their own internal processes yet. They are still charging you for human hours that have been commoditised by AI.
When you evaluate your next SaaS purchase, ask: Is this tool allowing me to bring a previously outsourced function in-house? If the answer is yes, the ROI isn't just the cost of the software; it’s the deletion of the agency fee.
Moving from Intention to Action
According to my data, while 73% of SMB owners say they want to use AI, only about 15% have actually integrated an AI-native tool into their core workflow. This 'Intention-Action Gap' is your biggest competitive opportunity.
Your competitors are likely still 'SaaS Taxidermists'—they are using old tools with new buttons. By asking these five questions, you ensure that you aren't just buying another subscription, but building a leaner, more autonomous business.
The window for this first-mover advantage is closing. Within two years, AI-native operations will be the baseline. Right now, it’s a superpower.
Where will you start? If you aren't sure, my recommendation is always to start where the volume is highest and the complexity is lowest. Usually, that’s your back-office operations or your first line of customer support.
Stop asking "if" and start asking "where". The answers are already in your spreadsheet.