A capability map for marketplace seller teams instead of a job forecast

A marketplace seller software skills forecast for England cannot credibly state how many jobs, vacancies or shortages will exist in 2027. No current official forecast reviewed on 6 September 2026 matches this exact occupation or buyer population. Plan capabilities against operational scenarios instead, and leave headcount and demand figures blank.

Use the available skills evidence narrowly

Observed current record, England. Skills England's AI foundation skills for work benchmark was published on 28 January 2026 and applies to England. It covers basic technical, non-technical, responsible and ethical capability for simple AI tools. It is not a forecast of marketplace roles or a certificate of competence.

Observed research, UK AI sector. The AI Labour Market Survey 2025, published by DSIT in January 2026, studies the UK AI labour market through surveys and interviews. Its population is not England marketplace sellers. Do not transfer its shortage findings or figures to this niche.

These sources justify checking AI literacy where an AI scenario exists. They do not decide staffing.

Build capability cards from the seller job

For each card, write the task, permitted decisions, evidence, assessor, exercise, acceptance field, expiry and fallback.

Product and offer evidence

The operator needs to trace identifiers, product facts, total price, optional choices and correction states. CMA price-transparency guidance supplies a current UK consumer reference. A qualified consumer reviewer owns legal interpretation; the software operator owns field provenance and reconciliation.

Safety and withdrawal

For goods, someone must distinguish seller, producer, importer, distributor and marketplace evidence, then stop or withdraw an affected offer. OPSS product-safety advice is a general official map. Product category and actor still need specialist review. A marketplace label cannot replace it.

Data and tracking

The team must map sources, purposes, controller and processor facts, device operations, consent or exception evidence, access, retention and deletion. The ICO's storage and access technologies guidance was finalised in April 2026. Competence should be assessed on the configured flow, not attendance at generic training.

Account security and recovery

Operators need least-privilege access, protected logs, incident escalation, credential removal, rollback, export and isolated restore. NCSC supplier-assurance questions can structure evidence requests, but answers and certificates remain bounded evidence. A cyber-security assessor should define and observe the recovery exercise.

AI supervision

Where the buyer proposes AI assistance, staff need to verify sources, recognise uncertainty, reject unsafe output and stop an out-of-scope action. NCSC's August 2026 agentic AI advice is interim and calls for oversight, monitoring, isolation and emergency shutdown. Recheck it when formal guidance appears.

Accessible operations

An accessibility practitioner should identify staff and customer tasks, barriers, adjustments and evidence. The EHRC services code of practice is current for services in England, Scotland and Wales. It does not supply a universal software test or job description.

Use scenarios to decide capacity

Buyer scenario: a new product-evidence field becomes necessary. Required capabilities are schema change control, source validation, safety review and rollback. Trigger and staff quantity: blank.

Buyer scenario: an AI assistant proposes listing changes. Required capabilities are provenance review, permission design, monitoring and emergency stop. Production authority remains disabled.

Buyer scenario: a supplier or marketplace interface changes. Required capabilities are export reconciliation, manual continuity, contract review and isolated restore.

For each scenario, record the volume and timing only from buyer evidence. Stop if no competent owner, independent reviewer or tested fallback exists.

Next step

Do not recruit against a speculative title. Test the capability cards with one fictional product-to-order fixture, record the evidence gaps and assign named reviewers. Recheck official sources before publication. Any later claim about jobs, shortages or salaries needs a current study whose population, geography and method actually match this niche.