95% of AI initiatives fail – that’s why Aixia delivers real value in production.
A new MIT report shows that 95% of companies’ generative AI initiatives fail to deliver business returns. This confirms something we’ve long seen in the field: technical demos are not enough. What matters is proven value in real workflows – in production. At Aixia, we take full responsibility from idea and PoC/PoV all the way to operation and support, without involving a plethora of intermediaries. The result is shorter time-to-value, lower total cost and solutions that are actually used.
What the MIT report shows
MIT Media Labs NANDA study The GenAI Divide: State of AI in Business 2025 finds that despite $30-40 billion in enterprise investment in GenAI (in the US), 95% of organizations see no business benefit. Only a small minority are getting their AI tools all the way into live operation with measurable Profit & Loss (P&L) impact – a clear “GenAI divide” between adoption and actual transformation.
MIT also summarizes where things often go wrong: pilots get stuck before production, tools lack memory/learning and are poorly integrated into everyday processes. So it’s not model ‘IQ’ that stops value, but the lack of learning systems that fit into real workflows.
MIT News further refers to the fact that only around 5% of AI pilots provide rapid revenue acceleration, while the majority stay at zero or near zero P&L impact. Read more about the report here: The Register
Aixia’s position: Lead with value – deliver in production
This is the core of Aixia’s way of working. We have completed a large number of AI projects in the Nordic region – from idea and PoC/PoV to production – where we ourselves drive the entire journey: architecture, development, integration, deployment and ongoing operation/support.
What it means for you as a customer:
- Total responsibility without unnecessary intermediaries. A single team to take you from PoC/PoV to production – faster, safer and cheaper than fragmented partner arrangements.
- Integration into your workflows. We design for real processes (ERP/MES/PLM/ITSM, data pipelines, security, governance) so that the solution is actually used – day one.
- Operation and support from the start. Our contracts and operational services are already in place at go-live, reducing the risk of ‘pilot-to-nothing’ and ensuring continuous improvement.
- Technology-agnostic delivery. We choose the right tool for the job – not the other way around – and can therefore optimize for time-to-value and total cost.
Proven in real life: automation that delivers real results
We have shown time and time again that AI-driven automation creates tangible value in real-world environments, such as:
- automated quality assessment/classification in industrial flows,
- intelligent document and image interpretation for faster processing,
- real-time support in production/maintenance that reduces errors and downtime,
- AI-assisted data center and IT operations that free up time and reduce costs.
What these projects have in common is measurable impact (time savings, quality, cost) and that they run in production, not in the lab.
Are you stuck with PoC – or unsure about its value?
The MIT report makes it clear: the market does not reward more experimentation. It rewards value in operation.
If you:
- is stuck with several PoCs that never get done,
- have doubts about whether AI can really create business value for you,
- need a concrete next step from PoV to production,
