Connecting Ideas, Coding The Future.

88% of organisations now use AI — so why do so few see real results?

Global surveys show AI is now mainstream, yet most organisations are stuck in pilots. Here is what the small group of high performers does differently — and how SMEs can copy it.

Artificial intelligence is no longer an experiment for a few tech giants. McKinsey’s global survey The State of AI in 2025 found that 88% of organisations now use AI in at least one business function, up from 78% a year earlier. Generative AI use has grown just as fast, and 62% of respondents say their organisations are at least experimenting with AI agents — systems that can plan and carry out multi-step tasks.

But the same research highlights an uncomfortable gap: most organisations are still in pilot mode. Only around a third report scaling AI across the business, and only a small group of “high performers” — roughly 6% of respondents — see a significant impact on profit.

What high performers do differently

According to the survey, the organisations getting real value from AI tend to:

  • Aim for growth, not just cost-cutting. They use AI to serve customers better and launch new services, not only to reduce headcount.
  • Redesign workflows. Instead of bolting AI onto old processes, they rethink how work flows end to end.
  • Have committed leadership. Senior leaders own the AI agenda and track results.
  • Manage risk deliberately. They put guardrails in place for accuracy, privacy and security.

Five practical AI projects for SMEs

You don’t need a data-science department to benefit. These projects work well for small and medium businesses in Rwanda and the region:

  1. A customer-service assistant that answers common questions on your website or WhatsApp — using only your approved information, and handing over to a person when it doesn’t know.
  2. Document and invoice processing — extracting data from receipts, invoices and forms to cut manual typing and errors.
  3. Marketing content support — drafting product descriptions, social posts and newsletters that your team then reviews.
  4. Sales and stock forecasting — spotting trends in your sales data to plan inventory and cash flow.
  5. Internal knowledge search — letting staff ask questions of your policies, price lists and procedures.

A simple roadmap

  • Week 1–2: pick one problem with a clear measure of success (e.g. “answer 60% of customer questions instantly”).
  • Week 3–6: build a focused pilot with real users and real data, and keep a human reviewing outputs.
  • Week 7–8: measure and decide. If it works, integrate it into your systems and train your team. If not, learn and adjust.
  • Then scale to the next process — this is where most organisations stall, so plan for it from the start.

Watch the risks

AI can be confidently wrong, can leak sensitive data if used carelessly, and can produce biased results. Keep personal data protected, review important outputs, and be transparent with customers when they are talking to an AI.

The lesson from the global data is clear: using AI is easy; getting value from it takes design and discipline. Start small, measure honestly and build on what works.

Want help choosing your first AI project? Request a quotation — we build AI assistants, automation and data dashboards for growing businesses.

Sources

Have an idea? Let’s build it together.

Tell us about your website, software or marketing goals and we’ll reply with a tailored proposal.

WhatsApp