TLDR: LONDON—Freshworks says UK mid market firms burn £11.7 billion a year on a complexity tax to correct noisy AI outputs, wasting about 25% of AI time and budget.
Key Takeaways:
- Freshworks frames the issue as an AI complexity tax created by legacy tech stacks and fragmented data, not bad intentions.
- UK mid market firms lose £11.7 billion yearly and 25% of AI budget to noise and errors that force workers to review and redo work.
- With only one in three using consistent AI governance and executives demanding returns within eight months, IT teams face rising career ROI pressure.
- Main fact: 1 out of every 4 hours spent correcting AI noise; 26% time lost to troubleshooting; 72% expect ROI within eight months.
AI is supposed to cut busywork, but many teams now do the busywork for the AI. Until governance and data quality catch up, every shortcut turns into a redo.
AI is supposed to cut busywork, but many teams now do the busywork for the AI. Until governance and data quality catch up, every shortcut turns into a redo.
Q&A
What changes when a company adds governance but does not fix its underlying data fragmentation?
Policies can document problems, but they cannot remove the root cause. Teams may follow rules while still spending hours correcting low quality outputs.
Why do AI complexity issues often show up as productivity loss instead of obvious system failures?
The AI looks usable at first, then quietly injects errors. Humans notice during review, so cost lands as rework and troubleshooting rather than outages.
What would a credible AI ROI plan look like for IT leaders facing an eight month return expectation?
They need measurable checkpoints tied to data quality, model performance, and reduced correction time, not just adoption metrics or tool spend.
How might workforce behavior evolve if employees learn that AI outputs usually require edits?
They may treat AI as a draft generator, which can preserve speed only if templates, guardrails, and domain specific datasets are in place.
Could the governance gap reflect a broader pattern seen in earlier tech rollouts like CRM and BI?
Yes. Many organizations moved fast on tools, then added process later. Freshworks suggests AI is repeating that cycle, with rework becoming the bill.
No comments yet. Be the first to share your thoughts!