The full insights are available exclusively to DataIQ clients.
Hire for skills, not increasingly fragmented job titles
Several peers said the proliferation of titles such as data product manager, data risk lead and enablement manager, is making recruitment harder. Similar work can be described very differently between organisations, while apparently similar titles can hide materially different responsibilities.
The response is to become much more explicit about the capabilities required. One organisation is using short but precise job descriptions which make it clear from the outset what is and what is not required in the role.
Another peer is shifting away from finding an exact title match and towards the underlying skills candidates can demonstrate: “I’m a big fan of the skill-based approach nowadays because of how diverse the jobs out there are.”
Rethink job architecture, not just hiring criteria
Several peers argued that organisations need to revisit how roles are defined and grouped. Emerging jobs such as data product managers, AI specialists and enablement roles often sit awkwardly within existing career frameworks.
Rather than continuously creating new job titles, some organisations are grouping roles according to transferable capabilities and shared outcomes. This makes it easier to recruit for skills, support internal mobility and give employees visibility of potential career moves.
Redesign recruitment for AI-assisted candidates
AI is weakening some traditional recruitment signals. Peers are seeing highly polished, increasingly similar CVs and suspect some candidates are using AI agents for real-time interview support.
One organisation replaced a lengthy coding-heavy first interview with shorter questions designed to test whether somebody was genuinely immersed in the work. Candidates were also shown deliberately flawed code — including AI-generated code — and asked how they would diagnose and improve it. Recruitment has also moved back to in-person assessment centres which also tests judgement and understanding over technical ability.
Use domain knowledge as an internal talent advantage
Several peers argued that deep organisational and business knowledge is becoming more valuable, not less. Internal candidates already understand processes, customers, stakeholders and organisational constraints; technical capability can often be layered on top.
“It is easier to teach the technical data and analytic skills than it is for someone to learn the business knowledge.”
Others are using internal projects, rotations, secondments and development programmes to help employees move into data and AI careers.
Put a premium on critical thinking in the AI era
As AI makes coding and technical execution faster, peers expect differentiation to shift towards judgement, curiosity and commercial thinking.
One leader described wanting people capable of acting as an “intellectual sparring partner to the AI” rather than accepting its recommendations. Others highlighted the need for practitioners who can understand the problem, communicate with colleagues and recognise when an apparently plausible technical answer is wrong.
AI may therefore reduce the value of some narrow technical tests while increasing the value of reasoning, challenge and business context.
Upskill through real work, not just more learning content
Peers were sceptical that large catalogues of e-learning alone will close capability gaps. Short 20–30 minute learning modules have a role, but several leaders are moving towards hands-on support around real business problems.
One organisation found considerably stronger engagement by working directly with teams on a live opportunity, helping them understand the data, tools and techniques required to deliver an outcome. Success then created a ripple effect as other teams asked for similar support.
Peers are also using champion networks to extend this model. One organisation is onboarding volunteers from individual functions with common AI guardrails and foundational knowledge, then relying on them to translate AI into the use cases and workflows that matter to their own teams.
Do not confuse training completion with capability
One peer found that fewer than a quarter of employees had participated in formal AI training, initially raising concerns among senior leadership. Yet usage data showed around 95% were already using the organisation’s AI tool daily.
That changed the diagnosis. The organisation did not need to force everybody through foundational training; many employees needed more advanced support, shared use cases and communities of practice instead.
Participation rates are therefore a weak proxy for capability. Usage patterns and observed behaviours can give leaders a better view of where intervention is actually needed.
Create time for learning rather than adding it to the day job
Strong executive sponsorship can accelerate AI adoption, but peers warned against making experimentation simply another expectation layered onto existing workloads.
One organisation is explicitly considering what work needs to stop or change so employees have time to learn. Leaders also need to accept a temporary productivity dip as people experiment with unfamiliar tools.
Without that trade-off, calls for continuous upskilling risk becoming rhetorical rather than operational.
The full insights are available exclusively to DataIQ clients.
