One of the most common conversations I have with businesses at the start of an AI engagement goes like this: they ask me what AI tools they should be using. I ask them what software they're already running. And somewhere in that conversation, it becomes clear that the AI capabilities they need are already sitting in tools they've been paying for, sometimes for years, and nobody has switched them on.
This isn't a criticism. The AI features built into mainstream business software have arrived fast, been added quietly, and are buried in interfaces designed for other purposes. But it does mean that for most SMEs, the first practical step on the AI journey doesn't involve buying anything new at all.
What's already in your existing software?
The AI features now embedded in mainstream business platforms cover an enormous amount of what SMEs need. Content generation, customer service automation, sales insights, financial forecasting, data analysis, and more. Here's a tour through the most commonly used platforms and what they now include.
HubSpot Breeze
HubSpot's AI layer, called Breeze, is now woven throughout the platform. It includes an AI content writer for marketing emails, landing pages, blog posts, and social content. It includes a prospecting agent that researches and personalises outreach automatically. It has a customer agent for handling support queries. And it offers predictive lead scoring that uses historical data to rank which prospects are most likely to convert.
If you're on any paid HubSpot plan, most of these features are available. Many businesses have them enabled without their marketing or sales teams knowing.
Microsoft 365 and Copilot
Microsoft Copilot is embedded across the entire Office suite. Word, Excel, PowerPoint, Outlook, and Teams. In Word it can draft, summarise, and rewrite. In Excel it can analyse data, spot trends, and generate formulas from plain English descriptions. In Outlook it summarises email threads and drafts responses. In Teams it transcribes meetings, generates action items, and produces summaries.
The full Copilot capability requires a Microsoft 365 Copilot licence, which is an additional cost. But the base AI features in the free and standard plans are already more capable than most businesses realise.
Salesforce Einstein
Einstein is Salesforce's AI layer, and it now includes generative AI for email writing, call summaries, case handling, and opportunity insights. For businesses with sales teams using Salesforce, Einstein can significantly reduce the time spent on CRM administration, drafting follow-up emails, summarising call notes, suggesting next actions on deals.
Xero
Xero has been adding AI-powered cash flow forecasting, invoice anomaly detection, and automated categorisation. These features reduce the time accountants and finance staff spend on routine processing and improve the accuracy of financial reporting. For SMEs who run their finances through Xero, the AI layer is often already active.
For the vast majority of SME use cases, the AI features built into existing business software cover the essentials. The Buy approach, using off-the-shelf AI in tools already in your stack, is the right starting point for 80% of SME AI initiatives. More complex integrations and custom builds come later, once the foundations are working.
Why these features often go unused
There are three reasons businesses don't use the AI features they're already paying for.
The first is awareness. New features appear in software interfaces without much fanfare, buried in menus or activated through settings pages most users never visit. Nobody sent a memo. Nobody ran a training session. The feature exists, and nobody knows.
The second is inertia. Even when people are aware of an AI feature, they default to the way they've always done things. Drafting an email the old way takes thirty seconds. Learning to use an AI draft effectively takes a little longer upfront, even if it saves time overall. Most people don't make that investment without encouragement.
The third is trust. Teams that haven't been trained to use AI output critically, to review it, edit it, and understand when it's wrong, are right to be cautious about relying on it. The answer to this is training and guidelines, not switching the feature off.
Don't confuse having AI features with using AI well. Turning features on without training the team how to use them responsibly, or establishing quality review processes, produces what some are calling the 'workslop' problem: a wave of mediocre, unreliable AI-generated content that creates more work than it saves. Switching it on is step one. Building the habits around it is steps two through ten.
How to take stock of what you already have
The practical first step is a software audit. For each major platform in your current stack, ask: does it have AI features? If so, are they enabled? Is anyone using them? What would it take to use them properly?
For most SMEs, this takes half a day and produces a clear picture of what's available, what's unused, and what would require a small investment of time and training to activate properly. It's one of the highest-ROI activities you can do early in an AI journey, because the costs are already being paid.
If you'd like a structured way to approach this, our AI Readiness Diagnostic covers the Tools & Technology dimension as part of a broader six-dimension assessment. The resulting report will tell you specifically where your current stack has untapped capability.
Take the free AI Readiness Diagnostic
See exactly where your business stands across the six dimensions of AI readiness. You get a personalised report back within one working day.
Take the diagnostic →