How Xtell's Data Works
This page explains exactly where Xtell's intelligence comes from, what it measures, and what its limitations are. We believe transparency about methodology is part of what makes intelligence trustworthy.
Live Vacancy Data
Source
Adzuna UK API
What it measures
Publicly advertised UK job postings, the same source used by the ONS for UK labour market monitoring. Vacancy volumes, salary ranges, and skills appearing in job advertisements.
What it does not measure
Internal promotions, word-of-mouth hiring, roles on company websites only, positions filled without advertising.
How to interpret it
Use as a demand signal showing employer intent. Trend direction is more reliable than absolute numbers.
ONS Payroll Employment
Source
ONS monthly payroll employment data
What it measures
Actual PAYE employment levels by sector, people in jobs, not job adverts.
How to interpret it
Compare alongside posting data: both rising = strong genuine demand; postings rising but payroll flat = high churn or slow actual hiring.
Refresh rate
Monthly (ONS release schedule)
Displacement Risk Scores - How They Are Generated
The research foundation
Xtell's displacement risk scores are informed by the body of published research on AI and the UK labour market, principally the UK Government AI Occupational Assessment published by the Department for Science, Innovation and Technology (DSIT) and the AI Security Institute in January 2026, the most comprehensive assessment of AI capabilities and UK labour market impact to date, published under Open Government Licence v3.0.
Key finding - UK exposure
70% of UK workers are in occupations where AI could perform or enhance their tasks, a higher share than the US or any other advanced economy.
Key finding - hiring impact
Hiring is falling faster in AI-exposed occupations, with UK job adverts falling 38% for high-exposure roles compared to 21% for low-exposure roles between 2022 and 2025.
How scores are currently calculated
Xtell's displacement risk scores are generated through AI-assisted task composition analysis, structured against the McKinsey Future of Work archetype framework and calibrated against the DSIT occupational exposure methodology. For each role, the scoring process assesses: the proportion of typical tasks that are automatable by current AI tools; which McKinsey archetype best describes how AI is changing the nature of work in that role (People-Centric, People-Agent, People-Robot, People-Agent-Robot, Agent-Centric, Robot-Centric, or Agent-Robot-Centric); and directional signals from UK vacancy data and published research on sector-specific AI adoption.
What this means in practice
Scores represent a structured, research-informed assessment of AI occupational exposure. They are directional intelligence, not actuarial precision. The direction of travel is clear for most roles; the precise timing and scale of impact in any individual role remains uncertain.
How scores are being refined over time
Xtell is designed to improve. Two refinement mechanisms are in development. First, live vacancy data integration: Adzuna UK vacancy trend data will be integrated to provide a real-time corroborating signal for each role, allowing scores to reflect actual hiring market conditions as they develop. Second, community validation: every displacement risk score on Xtell is open to validation by the professionals who work in these roles. Through the Community Intelligence feature, UK professionals can rate whether a score feels accurate, suggest an alternative, and classify their role against the McKinsey archetypes based on lived experience. Where community feedback diverges significantly from the published score, that role is flagged for review.
Why community validation matters
No algorithm, however well-designed, knows a role better than the professionals who work in it. Community validation is not a workaround; it is a deliberate and important part of how Xtell's intelligence matures.
How Seniority Affects Displacement Risk
Role-level scores
Displacement risk scores on Xtell represent a role-level assessment based on the typical task composition of that profession across all seniority levels.
Why seniority matters
In practice, seniority significantly affects individual risk within any role. Junior and mid-level professionals whose work is concentrated in transactional, volume, or process-driven tasks generally face higher displacement risk than the role-level score suggests. Senior professionals whose work centres on strategy, client relationships, complex judgment, and leadership generally face lower displacement risk than the role-level score suggests.
Personalised context
This is why Xtell provides seniority-specific context in the Role Evolution feature for every tracked role. The same displacement pressure that threatens a junior accountant doing compliance work is the same force creating more demand for senior accountants doing strategic advisory work.
How to interpret it
The displacement risk score is a starting point for understanding your situation, not a verdict on it.
AI sophistication and seniority
Research by KPMG and the University of Texas, published in Harvard Business Review in March 2026, analysed 1.4 million AI prompts from 2,500 employees over eight months and found that only around 5% of professionals use AI sophisticatedly. Employees above manager level were significantly more likely to use AI with deliberate strategy, for a greater diversity of tasks, and with clearer objectives — creating a compounding advantage that reduces their personal displacement risk beyond task composition alone. This distinction between AI comfort and AI sophistication helps explain why seniority reduces displacement risk in ways that go beyond task composition.
Task Composition and Displacement Risk
Overview
The displacement risk score reflects the average across all ways a role is performed. For many roles, actual risk varies significantly depending on the specific tasks a professional undertakes day-to-day.
How Xtell addresses this
Xtell provides task composition caveats on high-variance role pages — roles where the score could be materially higher or lower depending on task mix. For example, a Sales Development Representative doing primarily high-volume outbound email automation faces significantly higher displacement risk than one doing relationship-led enterprise prospecting. A Bid Manager focused on document production and compliance matrices faces higher risk than one leading win strategy and client intelligence.
How caveats are developed
These caveats are informed by published research on task-level AI exposure, professional community validation, and domain expertise. They are designed to complement the role-level score rather than replace it. Users are encouraged to use the community validation feature to indicate which task composition best describes their role.
The McKinsey Archetypes
Overview
Xtell classifies every role against one of seven McKinsey Future of Work archetypes. These describe how AI and automation are changing the balance between human workers, AI software agents, and physical robots in each role.
People-Centric
Humans handle most of the work, using AI as a supporting tool rather than a replacement. Human judgment, relationships, and presence remain central. Examples: administrative support, complex sales.
People–Agent
Humans work alongside AI software agents that provide real-time guidance, recommendations, or automation of specific subtasks. The human leads; the agent assists. Examples: customer service representatives, knowledge workers with AI copilots.
People–Robot
Humans work alongside physical robots that handle specific tasks, particularly repetitive or physically demanding work. The human provides oversight, judgment, and adaptability. Examples: maintenance technicians, hospital workers.
People–Agent–Robot
A mixed environment where humans manage both AI software agents and physical robots. The human role shifts toward coordination, oversight, and exception handling. Examples: surgeons using robotic assistance, advanced manufacturing operators.
Agent-Centric
AI agents handle most of the cognitive work. Humans act as supervisors who monitor outputs and intervene when needed. Displacement risk is high for those performing the tasks the agent now handles. Examples: automated content curation, automated customer service.
Robot-Centric
Robots perform most of the physical work. Humans manage, maintain, and oversee the technology rather than performing the tasks directly. Examples: automated production lines, warehouse picking.
Agent–Robot-Centric
AI agents and robots work together to complete tasks with minimal human intervention. Human roles are reduced to high-level oversight, system management, and edge-case resolution. Examples: fully automated supply chain fulfilment.
Closing note
Most roles sit clearly within one archetype today. As AI capability develops, roles may shift along the spectrum — from People-Centric toward Agent-Centric or Robot-Centric. The archetype assigned to each role on Xtell reflects the current trajectory, not a fixed permanent state.
Role Evolution Forecasts
Source
McKinsey Future of Work archetype framework applied to UK role data
What it measures
Projected transition in how work is divided between humans, AI agents, and automated systems based on current market signals and near-term trajectory.
What it does not measure
Guaranteed outcomes. Labour markets are inherently unpredictable.
How to interpret it
Use as directional intelligence to inform career planning, not as a definitive prediction.
Skills Intelligence
Source
LinkedIn Skills on the Rise 2026
What it measures
Skills appearing with increasing frequency in UK job postings, a forward signal of where employer demand is heading.
Refresh rate
Annual LinkedIn publication
Creative and Digital Roles - A Note on Methodology
Data challenge
Creative and digital professions present a different data challenge from knowledge work and trades roles. Many creative professionals work freelance or on short-term contracts rather than in salaried positions advertised on job boards. Standard vacancy data captures only part of this market, particularly for make-up artists, video editors, and copywriters where freelance work is significant.
How we address it
Displacement risk scores for creative roles reflect the best available combination of task composition analysis, industry body research, and directional vacancy data. Where vacancy data is limited, we will draw on industry body research to supplement the picture.
How to interpret it
Treat creative role displacement scores as directional rather than data-precise, and read the full role intelligence including what changes and what stays for the complete picture.
Supporting Research
King's College London, December 2025
Firms with workforces highly exposed to AI reduced total employment by 4.5% on average, with the effect concentrated almost entirely in junior positions, which fell by 5.8%.
Morgan Stanley, January 2026
UK firms reported net job losses of 8% over the past 12 months linked directly to AI, the highest of any country surveyed, including Germany, the US, Japan, and Australia.
Chartered Institute of Personnel and Development, November 2025
Junior roles stand to be most affected by AI, with early-career and lower-level workers across finance, insurance, IT, and admin bearing the brunt of AI-driven changes.
McKinsey, July 2025
Between 2022 and 2025, UK job adverts fell by 38% for high-exposure occupations compared to 21% for low-exposure roles.
Real-time monitoring
The UK Government's own January 2026 assessment concluded that real-time data monitoring of job postings in AI-exposed occupations is essential to track impacts as they develop and enable timely responses.
Xtell was built to do exactly this.
Important Caveats
Displacement risk scores reflect AI occupational exposure based on task composition analysis. They should be read alongside the seniority caveat displayed on every role card as risk varies significantly by seniority, sector, and specific responsibilities.
Scores reflect AI exposure as the primary factor. In practice, UK employer hiring decisions are also shaped by broader cost pressures, including the April 2025 employer National Insurance increases, National Living Wage rises, and Employment Rights Bill compliance costs. These factors are creating additional incentives to reduce headcount in AI-exposed roles independently of AI adoption itself. The combined effect on hiring is greater than AI exposure alone would suggest.
The direction of travel is clear. The precise timing and scale of impact in any individual role remains uncertain.
What Good Data Looks Like Here
No single data source tells the complete story of the UK labour market. Xtell combines published research, structured analysis, live vacancy signals, and increasingly direct input from the professionals most affected, to give the most complete and honest picture available.
All scores are directional. All forecasts are indicative. Labour markets change in ways that no model can fully predict.
That is why community validation exists. The professionals using Xtell are part of how it gets better.
Government Data Sources
Xtell integrates data from four UK Government sources to enrich role intelligence:
1. ONS Standard Occupational Classification (SOC 2020)
All Xtell roles are mapped to official ONS SOC codes, ensuring alignment with the UK's authoritative occupation taxonomy.
2. UK Government Occupations in Demand
Demand level classifications (Critical, Elevated, Not in High Demand) are sourced from the GOV.UK Occupations in Demand annual release, published by the Department for Education.
3. ONS Annual Survey of Hours and Earnings (ASHE)
Median salary data is sourced from ONS ASHE Table 14, providing independently verified wage benchmarks by occupation.
4. Skills England Occupational Maps
Role overviews and green job classifications are sourced from the Skills England public API.
All government data is used under the Open Government Licence v3.0. Contains public sector information licensed under the Open Government Licence v3.0.
Emerging Roles and SOC 2020 Classification Gaps
SOC 2020 was published before the current wave of AI deployment and does not include dedicated classification codes for many roles Xtell tracks, including Prompt Engineer, AI Ethics and Governance Lead, XR Developer, and other AI-adjacent and immersive technology roles.
Xplorient Limited submitted evidence to the ONS SOC 2030 public consultation in March 2026, calling for dedicated classification codes for emerging AI, XR, and technology roles. Our submission cited Xtell's vacancy monitoring data and the growing demand for roles that have no official SOC classification as evidence for new unit group creation.
Where a role has no SOC 2020 code, this is noted on the relevant role page. Xtell will adopt SOC 2030 codes when published and will release a full concordance table mapping Xtell roles between SOC 2020 and SOC 2030 classifications.
A Note on Physical Robotics
Xtell's displacement risk scores are primarily grounded in the UK Government's DSIT AI Occupational Assessment (January 2026), which measures the proportion of tasks within a role that AI and automation systems could perform. This assessment was designed to evaluate AI and cognitive task automation exposure. It was not designed to fully account for physical robotics displacement.
Physical robotics represents a distinct and evolving category of displacement risk. Where AI automates cognitive and administrative tasks, general-purpose humanoid robots and specialised physical automation systems threaten roles that involve physical presence, dexterity, and environmental adaptability, roles that current AI exposure assessments classify as low-risk.
For example, a Childminder scores 2% AI displacement risk because no AI system can replicate the physical presence, emotional attunement, and moment-to-moment judgment the role requires. This assessment is accurate for current AI systems. It does not account for the long-term trajectory of humanoid robotics platforms, which are advancing rapidly and may reach commercial deployment capability within a decade.
Roles most affected by this gap, where AI displacement risk is currently low but robotics exposure may be material over a longer horizon, include care and social roles, cleaning and domestic services, gardening and grounds maintenance, agricultural work, and some trades.
Xtell is actively working to incorporate physical robotics exposure as a distinct fourth dimension of the Role Intelligence Compass, alongside Displacement Risk, Extension Score, and Human Primacy Index. Until that work is complete, users of the platform particularly those in physically present roles should consider the longer-term robotics trajectory alongside the current AI displacement risk score.
All scores on Xtell are directional intelligence grounded in the best available government data. They are not actuarial predictions. The scoring methodology will be updated as the evidence base evolves.
Last methodology update: March 2026
The Xtell Role Intelligence Compass
Displacement risk is only one dimension of how AI affects professional roles. Xtell tracks two additional dimensions to give a complete picture:
Extension Score
How much AI amplifies the capability and productivity of skilled professionals in a role. A high extension score means professionals who embrace AI become significantly more capable, doing more, faster, to a higher standard. The Extension Score recognises that AI is not only a threat but a multiplier for human talent.
Human Primacy Index
How important it is that a role remains distinctly human, ethically, practically, or commercially. Some roles carry legal accountability, patient or client relationships, or creative authenticity that people specifically value as human. The Human Primacy Index recognises that technical capability to automate does not equal social or ethical appropriateness of automation.
All three scores are directional assessments, currently being refined through community validation by the professionals who work in these roles.
The Xtell Role Intelligence Compass and its component methodologies were developed by Xplorient Limited and first published in March 2026. The Extension Score and Human Primacy Index are proprietary to Xplorient Limited. Where this framework is referenced or cited, attribution to Xtell by Xplorient Limited is required. © 2026 Xplorient Limited.
Read the full framework explanation →