25 things we learned in 2025
2025 has been a year full of change. In technology, in the ways we work, in the shifting of industries and more.
QA is at the leading edge of constant transformation – and since we’re all about learning, we’re always on the lookout for what change and experience can teach us.
We’ve been keeping a finger on the pulse all year, and now we’re breaking it down>, so you don’t have to – playing back the hits of 2025 with 25 lessons learned, brought to you>r by our leading experts.
Read on for 2025’s most important insights on everything from harnessing AI, to the secret to successful diversity and inclusion, to how to hack distractions for better productivity…
1. Distraction is the enemy of productivity - but AI could be the hero
Did you know that, on average, we get distracted at work every 2.5 minutes? Worse, it can take around to refocus every time.
Here’s why:
-
117 daily emails
-
153 daily Teams messages
-
50 messages outside of core working hours
All those ‘pings’ are not just annoying, they’re a serious hit to productivity.
On our podcast, Simon Lambert, Microsoft UK’s Chief Learning Officer, explained:
“This is putting an extra pressure on, the knowledge worker to effectively get their work done. There needs to be a recognition of that, and we need to rethink how we work.”
But AI tools like Copilot can help – with users completing tasks 29% faster on average!
2. Jobs won’t all be replaced – but they will change
85% of jobs that will exist in 2030 haven’t yet been invented. That doesn’t have to mean widespread employment panic, but it does call for adaptability at all levels – new skills in line with emerging tech will strengthen careers.
3. Skills can hide where you least expect them
If you’re considering, as a business, investing heavily in AI tooling, you need to know who will deploy it.
4. Skills still matter; AI cannot replace Eengineers
As Andy Smith, Portfolio Director for Software and DevOps>, , AI cannot replace all skills.
“Anecdotal stories are now being confirmed by empirical data, outlining that 45% of AI-generated code samples introduce known security vulnerabilities.”
Organisations must continue to prioritise and develop talent, or face fatally flawed outcomes further down the line.
5. Start small with data transformation – Big Bang approaches fail
When it comes to data and AI adoption, many organisations still go wrong by trying to do too much, too fast.
how to avoid this recipe for failure:
“Start small, pilot, run a proof of concept and do not undertake a Big Bang approach. Scope small projects, test them, trial them, educate people.”
The lesson? Incremental delivery and cultural readiness beat grand plans every time.
6. Digital identity schemes demand s trust, not just tech
The UK government has recently brought the debate on digital identity back to the fore. But why hasn’t it worked before?
explained that failed systems weren’t broken by code - they collapsed due to poor design and lack of public confidence.
“Back in the early 2000s a similar scheme was introduced under Tony Blair and later scrapped, having failed to gain the public trust it needed. People’s mental models of something like this are fixed in the past.”
7. Bias starts with data - and ends in exclusion
Did you know that misidentify dark-skinned women 34% more often than light-skinned men? This is just one example of the importance of diversity and inclusion in technology and product design.
, highlighted how better understanding through data leads to better outcomes.
“Until you put the right data into the hands of the decision makers that matter, you're not going to shift the system. Without solving the data challenge, we can't solve the rest of the challenges. "
8. Cybersecurity starts at the helpdesk
Cyber breeches seem to have become a regular occurrence, with increasing impact - and even the biggest organisations can fall victim. But is the root cause really flawed tech?
revealed that undertrained frontline staff are often the weakest link in cyber defence. Capability-building must start at the coalface, and security is everyone’s responsibility.
“The data that that helpdesk teams work with, and the data they generate, has been underutilised. One reason why they are at the forefront is because they work with end users every day.”
9. The simplest features can be the most impactful
, and our podcast host Paddy Dhanda>, discussed a few of their favourite and most used tech features:
-
Copy and paste! Classic for a reason. Just imagine trying to live without it…
-
AI transcription, for powerful timesaving and insight extraction.
-
Scheduling messages, to respect the varying schedules of global teams.
-
Automatic subtitles when rewinding on streaming services – showing that the teams behind the tech understand user behaviour; you’re jumping back ten seconds because you didn’t hear what someone said.
A theme emerges here. Sometimes great user experiences don’t come from complexity, but from simplicity itself, and product design that truly understands user needs and pains.
10. Lifelong learning is leadership hygiene
How do senior leaders still fall behind in digital skills?
Jo Bishenden argues it’s because learning is still seen as optional at senior levels:
“If your boss understood less than you about the tools and technology you work with – and more importantly, their context and impact – would it fill you with trust in their strategic direction?”
Reskilling every five years is now the minimum. Leadership isn’t about having all the answers – it’s about asking better questions, staying informed, and modelling curiosity.
11. Prompt literacy is now a core skill
AI is now a ubiquitous and essential tool – so knowing how to ask the right questions of AI is becoming essential for digital workers.
in the US, 85% say that prompting will be a vital skill by 2030, but right now only 31% of employers offer training in this area – a gap that needs to be closed.
12. AI adoptions demand a balance of speed and safety
AI offers huge opportunities, but also introduces new risks. Security must evolve in step with new capabilities.
Organisations know this, with citing AI risk and compliance as a key barrier to scaling, while only 8% have governance fully embedded. The gap is costing businesses time, confidence, and competitive advantage.
13. Fairness isn’t just statistical – it’s human
The 2020 A-level grading fiasco showed what happens when algorithms ignore human expectations.
Philosopher of technology Tom Chatfield explained:
“Fairness is not a product of algorithms on their own… It requires accountability, transparency, and mechanisms for appeal.”
For leaders deploying AI, there’s a clear lesson to learn - optimise for trust, not just efficiency, because user and audience buy-in will make or break your solution.
14.Data maturity is the new competitive advantage
Why do so many organisations fail to get value from their data?
Amma Ainsley explained that it’s not just about tools - it’s about culture, skills, and governance:
“If you don’t move quick, if you haven’t got the right people with the right competencies, the right strategy… you’re not going to compete.”
Data literacy and incremental change are now core leadership priorities.
15.Leadership and culture – not just investment – drive successful tech adoption
Why do so many transformation programmes fail, even with the best technology?
Our portfolio director for (and podcast host!), Paddy Dhanda, explained that success depends on people, not just platforms:
“One of the critical skills for tech leaders in the modern world is their ability to communicate and inspire their people.”
But inspiration alone isn’t enough:
“You can have the best technology, but culturally you might have a very resistant employee base.”Clear vision, strong leadership, and a culture that embraces change are the real foundations of digital transformation.
16. AI empowerment, not replacement
why AI won’t replace project managers – but the time to upskill is now:
“AI isn’t here to take your job – it’s here to make you better at it.” Those who embrace AI-driven tools like predictive analytics will move from task execution to strategic leadership.
17. The end of single-platform dominance
why we may be facing the end of single-platform dominance – and what that means for IT leaders:
Windows is no longer the default. macOS, Linux, iOS, and Android are now critical. Unified endpoint management is essential for security and agility.
18. Multi-region resilience beats single-cloud hubris
Stuart Scott warned in :
“Cloud has become shorthand for resilience… but it’s something you architect.”
A single-region outage can cripple operations. Multi-region strategies are now non-negotiable.
19. Human empathy is the new industrial edge
Steve Rouse argued in :
“As machines get better at being machines, humans have to get better at being more human.”
Empathy and emotional intelligence will define competitive advantage in an AI-driven world.
20. Apps may soon run inside AI, not the other way round
Dr Vicky Crockett explained in
“When ChatGPT becomes the ‘interface’ for the internet… if your app isn’t in ChatGPT, you just became invisible.”
What does all this mean? Prompting is the new UX, and businesses will need to integrate with AI in order to reach their users.
21. Hyperscalers are still indispensable
Stuart Scott highlighted in :
“Only hyperscalers have the capital and infrastructure to deliver… frontier AI.”
NEO clouds bring flexibility, but hyperscalers remain the backbone of global AI scale.
22. Cyber espionage is harder than ever to detect
Richard Beck revealed in his blog, , that “adversaries can now silently rewrite your systems, leaving without a trace.”
Cyber defence must shift from perimeter security to detecting stealthy, internal compromises.
23. Skills-first is the new L&D
According to Luke Radford in , “Businesses need skills at pace… the half-life of skills is shrinking.”
The lesson is that L&D must move from static role-based training to dynamic, AI-powered skill ecosystems to keep in step with change.
24. AI usage and governance aren’t adding up
How many organisations are using AI? >.
But how many actually have governance fully embedded? !
True accountability requires legal, compliance, and operations working together. That’s how you build safe, successful and sustainable AI rollouts.
25. Cloud underpins AI explosion
As Stuart Scott explained in
“AI can’t scale without cloud.” That’s why we’re seeing a trend of huge cloud compute deals between the likes of Oracle and OpenAI, or Meta and Google Cloud.
It’s a reminder that in the age of AI, infrastructure is a deciding factor that can spell success or failure.