If you run or lead a business with around 100 employees, you have probably been told that AI will transform everything. The presentations are impressive. The demos are slick. The promises are enormous.
But when you get back to your desk and look at your actual business — your team, your processes, your budget — the gap between the hype and reality feels vast.
Here is the honest picture.
Where AI genuinely helps right now
Let us start with what actually works for businesses your size, today, without a research lab or a seven-figure budget.
Document processing and data extraction
This is the most reliable win for medium-sized businesses. If your team spends hours reading documents, extracting information, and entering it into systems, AI can handle a significant portion of that work.
Think invoices, contracts, compliance documents, application forms, survey responses. AI does not just read faster — it reads consistently. No Friday afternoon mistakes. No missed fields. No backlogs after holiday cover.
Realistic impact: 60-80% reduction in manual processing time for structured documents. Less for unstructured content, but still meaningful.
Customer inquiry triage and routing
Not a chatbot that annoys your customers. A system that reads incoming emails, tickets, or form submissions and routes them to the right person with the right context already attached.
Your customer service team still handles the conversations. But instead of spending the first few minutes of every interaction working out what the problem is and who should deal with it, they can start solving immediately.
Realistic impact: 30-50% reduction in first-response time. Fewer misrouted inquiries. Better data on what your customers actually ask about.
Internal knowledge retrieval
Every growing business has the same problem: critical knowledge lives in people’s heads, in scattered documents, and in email threads from three years ago. AI can make that knowledge searchable and accessible.
This is not about replacing your experienced staff. It is about making sure their knowledge is available when they are on holiday, in a meeting, or have moved on. New starters get up to speed faster. Decisions get made with better information.
Realistic impact: Hard to quantify precisely, but businesses consistently report faster onboarding, fewer repeated questions to senior staff, and better decision-making quality.
Report generation and data summarisation
If someone in your business spends half a day every week compiling a report from multiple data sources, AI can probably do the compilation in minutes. Your team then spends their time analysing and acting on the data rather than gathering it.
Realistic impact: 70-90% reduction in report preparation time. Better consistency across reporting periods.
Where AI is not worth it yet
This is the part most consultants skip. But it matters, because wasting budget on the wrong AI project can set back your entire AI strategy.
Replacing skilled judgment calls
AI cannot replace your experienced team’s judgment on complex, context-dependent decisions. It can provide data to inform those decisions, but the actual decision-making — especially where relationships, organisational politics, or nuanced risk assessment are involved — still needs human expertise.
Small-volume processes
If a task takes one person two hours a week, automating it with AI might cost more than the time it saves. AI projects have setup costs, maintenance requirements, and learning curves. The economics only work when the volume justifies the investment.
Customer-facing interactions without a safety net
Letting AI handle customer conversations directly — without human oversight — is still risky for most medium-sized businesses. One badly handled interaction with a key client can cost more than years of efficiency savings. Start with AI-assisted, not AI-replaced.
The honest starting point
The businesses that get the most from AI are not the ones that start with the biggest project. They are the ones that start with the clearest problem.
Pick one process that meets these criteria:
- It consumes significant staff time (at least 10-15 hours per week across your team)
- It follows relatively consistent rules and patterns
- The data involved is already digital (or easily digitised)
- A mistake is recoverable (not a compliance disaster or a lost client)
- You can measure the improvement clearly
Get that working. Prove the value. Build internal confidence and capability. Then expand.
That is not a timid approach. It is the approach that actually works.
What this means for your budget
A focused first AI project for a 100-person business typically costs between £8,000 and £25,000, depending on complexity. That should include proper scoping, implementation, testing, and training for your team.
If someone is quoting you six figures for a first project, they are either overcomplicating it or they are selling enterprise solutions to a medium-sized business. Neither is in your interest.
The ongoing costs — tool licenses, API fees, occasional maintenance — usually run £1,000 to £3,000 per month for a well-scoped project.
Most businesses see positive ROI within three to six months on a well-chosen first project. Not because AI is magic, but because the right problems are genuinely expensive to solve with human time alone.
Next steps
If you are thinking about where AI could help your business, the worst thing you can do is buy a tool and hope for the best. The best thing you can do is get specific about your problems before you start shopping for solutions.
That is exactly what an AI Clarity Session is designed for — 90 minutes to map out your actual opportunities, with realistic costs and timelines, and no obligation to proceed.