Artificial intelligence is now part of everyday work. Employees are using AI to draft emails, analyse data, automate repetitive tasks, and improve productivity. As adoption accelerates, it's easy to assume that confidence with AI is keeping pace with capability.
But is it?
Kaplan's ‘From AI to ROI: Who owns the problem?’ report found that 80% of employees say they feel confident using AI. HR/L&D leaders estimate that, on average, 37% of their employees are proficient. C-suite leaders estimate 46%.
It's a finding that raises an important question: are organisations confusing confidence with capability?
Many employees are comfortable experimenting with AI tools. But feeling confident using AI isn't the same as having the skills to apply it effectively and consistently in the workplace.
That distinction could be one of the biggest barriers preventing organisations from realising the full value of their AI investment. It could also create a risk for organisations if employees feel confident using AI without having the level of proficiency needed to use it effectively, potentially leading to misplaced confidence in how AI is being applied.
Confidence and capability aren't the same thing
As AI becomes more accessible, it's natural for employees to feel increasingly confident using it, with many now relying on AI to support everyday tasks.
That confidence is a positive sign, showing that employees are willing to embrace new technology rather than avoid it.
However, confidence alone doesn't guarantee capability.
In simple terms:
- Confidence is about feeling comfortable using AI.
- Capability is about using AI effectively to achieve better outcomes.
An AI-capable employee doesn't just know how to use AI tools, they can also:
- identify where AI can add value in their role and support business objectives
- use effective prompts and context to get useful, relevant outputs from AI tools
- critically evaluate AI-generated outputs, checking accuracy, relevance, and potential risks
- use AI responsibly, including understanding data security, privacy, ethics, and governance
- recognise when human judgement, expertise, and oversight are needed
Put simply, confidence helps people start using AI, whilst capability helps organisations generate lasting value from it.
Consider two employees using AI to draft a customer proposal. Both feel confident using the technology. One accepts the first response and sends it with only minor edits. The other refines their prompts, checks the accuracy of the information, adapts the tone for the client, and reviews the content for sensitive or confidential information before sharing it.
Both employees are confident using AI, but only one is demonstrating genuine AI capability. The difference isn't whether they use AI, it's how effectively they apply it.
Why do employees and leaders see AI capability differently?
The gap between employee confidence and leaders' perceptions of capability doesn't necessarily mean one group is wrong. Instead, it reflects the different ways they measure success.
Employees tend to assess their own experience. If AI helps them complete tasks more efficiently or makes parts of their job easier, they're likely to feel confident using it.
Leaders, however, are looking beyond individual experiences and are asking whether AI is improving productivity across teams and delivering measurable business value.
The wider findings from Kaplan's AI Skills Report help explain why these perspectives differ.
The report found that:
- 45% of HR and learning leaders say AI tools are being underutilised by the teams expected to use them.
- 33% of employees say they've been asked to use AI without sufficient training.
- 26% remain unclear about how AI should be applied within their role.
Taken together, these findings suggest the issue isn't a lack of enthusiasm. Instead, many organisations haven't yet provided the practical learning, guidance, or role-specific support needed to turn confidence into genuine capability.
This reflects what we see when working with organisations developing AI skills. Employees are often comfortable using AI for everyday tasks, however, far fewer consistently evaluate AI-generated outputs or understand the governance and security considerations that come with using it responsibly.
Why does the capability gap matter?
When organisations mistake confidence for capability, they risk overestimating how prepared their workforce really is.
Without the right skills, employees may struggle to:
- identify the best use cases for AI
- question inaccurate or biased outputs
- use AI consistently across different tasks
- maximise the productivity benefits AI can offer
The consequences can extend beyond day-to-day inefficiencies. Capability gaps can affect an organisation’s ability to realise the value of its AI investments, with 37% of C-suite leaders saying their organisation has failed to achieve the expected ROI from an AI tool or deployment.
This is largely due to:
- inconsistent adoption across teams
- underused AI tools
- governance and compliance risks
- missed opportunities to improve productivity
- slower progress towards wider digital transformation goals
Businesses may have invested in the technology, but without the capability to use it effectively, much of its potential remains untapped.
What does an AI-capable workforce look like?
Building AI capability is about helping people develop the practical skills needed to use AI effectively within their own role.
An AI-capable workforce is one where employees:
- understand where AI genuinely adds value
- use AI to solve real business challenges
- evaluate AI outputs rather than accepting them at face value
- work responsibly within governance and security guidelines
- continue developing their skills as AI evolves
These behaviours help organisations move beyond experimentation and towards meaningful, sustainable adoption.
How can organisations close the capability gap?
Closing the gap starts with recognising that AI capability can't be measured by confidence alone.
Rather than relying on self-assessments, organisations should focus on how employees apply AI in their day-to-day work and create learning that supports practical application.
This could include:
- assessing capability through real workplace outcomes rather than confidence levels
- providing role-specific AI learning tailored to different teams
- using apprenticeships and structured learning programmes to build AI skills alongside practical workplace experience
- giving employees opportunities to practise using AI in everyday work
- embedding responsible AI principles into learning programmes
- supporting continuous development as AI technologies evolve
Through investing in practical, ongoing learning, organisations can transform employee confidence into genuine capability and unlock greater value from their AI investment.
Final thoughts
As AI becomes increasingly embedded in the workplace, confidence is an important first step, but it shouldn't be mistaken for capability.
If four in five employees feel confident but fewer than half of leaders believe they're proficient, the challenge isn't getting people to try AI. It's helping them use it well.
Organisations that close that gap will be the ones that see the greatest return from AI, not because they adopted the technology first, but because they invested in the people using it.
As Kaplan's AI Skills Report concludes:
"The distinction between confidence and capability matters... Until organisations close that gap, AI investment will keep delivering mixed results."
Read Kaplan's AI to ROI report to learn more.
FAQs
What is the difference between AI confidence and AI capability?
AI confidence refers to how comfortable employees feel using AI tools. AI capability is their ability to apply AI effectively, responsibly, and appropriately within their role to achieve meaningful business outcomes.
Why do employees and leaders have different views of AI capability?
Employees often judge their confidence based on their own experience using AI, while leaders assess whether AI is delivering measurable improvements across the organisation. These different perspectives can create a gap between perceived confidence and observed capability.
How can organisations assess AI capability?
Rather than relying solely on confidence levels, organisations should evaluate how employees apply AI in their day-to-day work, whether they understand responsible AI practices, and how AI is contributing to business outcomes.
What skills do employees need to use AI effectively?
Employees need practical skills such as writing effective prompts, critically evaluating AI-generated outputs, understanding responsible AI use, and knowing when human judgement is required alongside AI.
How can organisations close the AI capability gap?
Organisations can build AI capability through practical, role-specific learning, apprenticeships, opportunities to apply AI in real workplace scenarios, ongoing development, and clear guidance around responsible AI use.
Methodology
The research was conducted by Censuswide, among three samples of:
- 1000 UK employees within transformation-heavy sectors and functions in companies with 50+ employees where AI deployment is active (excludes Business owners and C-Level Execs) between 10.07.2026 - 21.07.2026
- 600 UK HR and L&D decision makers (HR directors, L&D managers, People and Talent leads) with 50+ employees, whose organisation is using AI between 10.07.2026 – 22.07.2026
- 600 (Aged 25+) C-Suite and Director level; in specific sectors; whose organisations use AI; and at companies of 50+ employees between 10.07.2026 – 22.07.2026
Censuswide is a member of the Market Research Society (MRS) and the British Polling Council (BPC), and a signatory of the Global Data Quality Pledge. We adhere to the MRS Code of Conduct and ESOMAR principles