AI · · 3 min read
The AI tools audit — how to figure out which AI tools are actually worth paying for
Most businesses are either under-invested in AI tools or paying for six tools that overlap. Here's how to audit what you have, identify the gaps, and cut what isn't working.
By Mediseo

The typical business in 2026 is paying for AI tools. Maybe it's ChatGPT Plus for a few team members, Jasper or Copy.ai for marketing, an AI meeting summariser, a code assistant, and some AI features bundled into the project management tool.
The total spend is €300–800 per month. The actual productivity improvement is unclear. When you ask team members what they use consistently and what genuinely saves time, the answers are often surprising — not the tools you'd expect.
An AI tools audit is the process of answering three questions honestly: What are you paying for? What are you actually using? What's it worth?
Step 1: Inventory everything
Pull the credit card statement and the subscriptions list. Include:
- Standalone AI tool subscriptions
- AI add-ons within existing tools (Notion AI, HubSpot AI features, Salesforce Einstein)
- Per-seat licenses (ChatGPT Teams, GitHub Copilot)
- API costs if your team is using OpenAI or Anthropic APIs directly
For each tool, record: monthly cost, number of seats, which team members use it, and what they use it for.
Step 2: Audit actual usage
For each tool, get honest data on how often it's used and by whom.
Ask team members to log AI tool usage for one week — not theoretically, but actually. Which tools did they open? Which ones did they use to complete a task that would otherwise have taken longer? Which ones did they open, realise weren't helpful, and close?
Most audits reveal that 70–80% of usage comes from 1–2 tools. The rest are aspirational subscriptions — bought because they seemed useful, retained because no one got around to cancelling.
Step 3: Calculate value per tool
For the tools that get real usage, estimate what they're replacing:
- Time saved per week × average hourly cost of the person using it
- Tasks that wouldn't have been possible otherwise (content that now gets produced, analyses that now get done)
- Quality improvements that translate to business outcomes (better ad copy → higher CTR → lower CPA)
A writing assistant that saves a marketing manager 3 hours per week is worth €100–150/month at an average salary. If it costs €30/month, it's a clear keep. If it costs €200/month for features that only matter for a content team of 20, reconsider.
The overlap problem
The biggest waste we see is tool overlap. Three separate tools all doing variations of "AI writing assistant." Two tools both summarising meetings. An AI feature in the project management tool that does the same thing as the separate AI tool someone added last quarter.
Map the function of each tool (draft generation, meeting notes, code completion, data analysis, research) and look for redundancy. Usually there are 2–3 clear candidates for cancellation without any loss of capability.
What to keep, what to cut
Keep: Tools with consistent daily/weekly usage by multiple team members, tools where the time saving is clear and measurable, tools that enable tasks that otherwise wouldn't happen, tools with no equivalent in your existing stack.
Cut: Tools that only one person uses and that person can't articulate the value, tools that overlap with something else you're paying for, AI features bundled into platforms you're not otherwise getting value from, tools bought for a specific project that never got implemented.
Trial and kill: Some tools look good in demos and don't survive contact with real work. Set a 30-day trial with specific success criteria before committing to a subscription.
The deeper question
An AI tools audit often reveals that the problem isn't the tools — it's that no one has designed workflows that use them consistently. The ChatGPT subscription isn't delivering ROI not because ChatGPT isn't useful, but because people use it ad hoc rather than as a systematic part of defined processes.
This is where AI implementation — designing the actual workflows, not just buying the tools — makes the difference. Our AI implementation service addresses this: we help businesses get structured, measurable value from AI rather than a collection of subscriptions with unclear ROI. If you want to start with an honest audit, book a call.