TLDR: Swiss companies report trouble hiring artificial intelligence specialists while the same openings attract large applicant pools, so employers who define the job precisely and screen on evidence fill these roles fastest.
A third of Swiss companies report difficulty hiring artificial intelligence specialists, the same share that struggles with skilled hiring overall
The starting figure comes from the job platform operator JobCloud. Its jobs.ch KI Report 2026, released in June 2026, surveyed about 850 Swiss companies and 3,600 employees alongside 7.3 million job ads published on jobs.ch, jobup.ch and JobScout24.ch between 2019 and 2025. On the company side, 34% said they have difficulty recruiting suitable artificial intelligence (AI) specialists, and 57% said the question is irrelevant to their firm. Read together, the figures suggest that difficulty is the common experience among firms actually recruiting AI profiles. The report calls its surveys a snapshot of sentiment, so the 34% works best as a signal of direction.
The official statistics put that signal in proportion. The Federal Statistical Office (FSO) reported in its employment release of 27 August 2026 that 33.9% of Swiss companies, weighted by employment, had difficulty recruiting qualified staff in the second quarter of 2026, up 0.4 percentage points on the previous quarter. Firms reported 98,700 open positions, 2,600 more than a year earlier, and total employment rose 2.0% year on year. The two surveys measure different populations with different methods, yet the near-identical shares carry a useful message for hiring managers: as of September 2026, AI hiring in Switzerland is about as hard as skilled hiring in general. In Edward Galle’s view, ordinary recruiting discipline, precise ads, evidence screens and a fast process solve most of that difficulty.
| Indicator | Figure |
|---|---|
| Firms struggling to recruit AI specialists (jobs.ch survey) | 34% |
| Firms with difficulty recruiting qualified staff, second quarter 2026 (FSO) | 33.9% |
| Firms calling the AI recruiting question irrelevant to them | 57% |
| Scarcity index, AI-exposed occupations, 2023 to 2025 | −69% |
| Scarcity index, less AI-exposed occupations, 2023 to 2025 | +19% |
| Employees answering “yes, certainly” to feeling equipped for AI | 33% |
The shortage is a matching problem, because AI openings already attract hundreds of applicants
The Lucerne University of Applied Sciences and Arts (HSLU) found in its Swiss AI Jobs Report 2025 that AI roles attract hundreds or even thousands of applications, and Edward Galle’s analysis of the 4,703 HSLU job ads, Data Scientists, Engineers and Analysts Fill 43% of Swiss AI Job Ads, sets out the hiring-speed and selection-process detail behind that finding. The jobs.ch data points the same way: its scarcity index for AI-exposed occupations fell 69% between 2023 and 2025, while scarcity for less AI-exposed occupations rose 19%. In that index, AI-exposed means commercial occupations plus information technology (IT) and telecoms; the less exposed group covers roles such as nursing and the trades.
A firm that receives hundreds of applications and still reports difficulty is facing a matching problem. Three causes recur: the job ad describes a generic AI profile while the work requires a specific one, so the pile fills with the wrong people. The screen filters on job titles and degrees, so capable people with adjacent titles drop out early. And, in Edward Galle’s experience, a slow process loses the strongest applicants to faster employers before the final round.
Domain context sharpens the mismatch in pharma, medical technology (MedTech) and banking, where a data hire works inside good practice (GxP) quality systems, Swiss Financial Market Supervisory Authority (FINMA) supervision or medical-device rules, so a data engineer from e-commerce and one from a validated laboratory are different hires. In Edward Galle’s view, the regulated employer gains most by naming that context in the first lines of the ad: “validated environment”, “audit trail” and “change control” attract the candidates who already work that way and signal the difference to everyone else.
Job ads work when they say whether the hire will use AI or build it
Most Swiss AI demand sits in roles that apply AI tools inside an existing function, a pattern Edward Galle examined earlier in Swiss Job Ads Added About 38 Artificial Intelligence User Roles for Each New Builder Role in 2025. The first decision in any AI job ad is therefore which of the two the firm is hiring. A builder ad names the stack, the data and the deployment target: the cloud platform, the pipeline tools, the model types and the systems the work feeds. A user ad names the business process, the tools already licensed and the decisions the person will improve, such as forecasting demand, drafting medical information responses or segmenting customer relationship management (CRM) data.
Language and seniority deserve the same precision: a French or German requirement narrows the national pool sharply, so it belongs in the ad where the daily work demands it, for example presenting results to a French-speaking quality unit. Seniority also needs to be stated honestly, because an ad that asks for a senior profile at a junior budget generates exactly the volume-and-mismatch pattern described above.
A practical test before publishing: a strong candidate should be able to read the ad and name the first deliverable of their first ninety days. If they can, the ad attracts self-selected applicants; a list of AI buzzwords attracts everyone.
Skills-based screening finds capable people that title filters reject
Screening on titles fails in a field where titles move faster than work, and SwissICT’s Urs Gurtner-Oesch summed up the logic of its 2026 salary study with the line “Entscheidend ist, auf welchen Skills die Erfahrung beruht”: what counts is the skills on which the experience rests.
A skills-based screen turns that principle into steps. The hiring team lists three to five capabilities the role needs on day one, for example building a reproducible pipeline in Structured Query Language (SQL) and Python, documenting a model for an auditor, or explaining a result to a non-technical brand team. Each capability gets one piece of evidence: a repository, a validation summary or a short written case. The automated filter then checks for evidence over exact titles, and the technical case study replaces one of the later interview rounds. Edward Galle’s time-to-hire benchmarks for regulated roles give a realistic outer limit for the whole process.
For regulated employers, Edward Galle recommends one case study that mirrors real controlled work: give the candidate a short, anonymised change request for a data pipeline and ask for the impact assessment and test plan. The exercise takes about an hour, shows whether the person thinks in validation terms, and gives the quality unit a concrete artefact to review alongside the technical team.
Pay competition now runs on the expert track, where specialists out-earn middle managers
Swiss technology pay has been almost flat. SwissICT’s salary study 2025, published on 2 September 2025 from 39,511 salary data points at 231 organisations, found median base pay up 0.4%, with the most common senior level at 130,060 Swiss francs (CHF). Its 2026 edition, published by SwissICT on 25 August 2026 from 37,112 salaries at 214 companies, found overall median growth of 0.2%, yet Senior Experts reached a median of CHF 177,938, up 2.8% in a year and 16.2% since 2020, against CHF 156,000 for middle management. The study names AI, cloud, cybersecurity and data as the skills in highest demand.
| Track (SwissICT 2026) | Pay in 2026 | Growth |
|---|---|---|
| Senior Experts | CHF 177,938 (median) | +2.8% on 2025; +16.2% since 2020 |
| Middle management | CHF 156,000 (median) | n/a |
| Upper management | n/a | +17.9% since 2020 |
| Professionals | n/a | +1.9% since 2020 |
| Juniors | n/a | +1.3% since 2020 |
| Junior new hires | CHF 80,000 (average) | +2.6% on 2025 |
A senior data or AI specialist compares offers against an expert track that already pays above a management role, so a team-lead title used as the main lure misreads the market. Firms that win these hires offer a defined expert career path, with pay bands that rise on capability, and reserve management roles for people who want to manage. Junior new hires, at an average of CHF 80,000 in 2026, leave room to grow talent internally at a fraction of the expert-track cost.
In pharma and MedTech, the expert track has a natural home that many companies already run: the subject-matter expert who signs off validation, owns a data standard or represents the function in inspections. Edward Galle’s view is that AI and data specialists fit that existing ladder well, because the ladder already rewards depth and accountability, and extending it to data roles costs less than inventing a new career framework.
Zurich still leads AI hiring, and Geneva posted the largest gain in share
In the jobs.ch report, Zurich’s share of national AI-related ads fell from about 38% before the breakthrough of large language models to about 32%, because other regions grew faster while Zurich kept adding roles. Geneva recorded the strongest absolute gain in its share of national AI job ads, and the report names Basel and Vaud as regions where life sciences and MedTech clusters create roles that combine medicine, data analysis and AI.
Employers in Geneva, Lausanne, Basel and Bern compete with Zurich on three factors. The work itself comes first: a role close to a molecule, a device or a patient journey carries real meaning, and the ad should lead with it. Flexibility comes second: SwissICT’s 2025 study found that over 94% of participating organisations offer home-office days, while the median requirement for office presence rose to 2.5 days a week, so a firm that sets presence by team need has a visible advantage. Reach across borders comes third: cross-border commuters from France around Geneva and Basel hold G permits under free movement, and European Union (EU) and European Free Trade Association (EFTA) nationals hired directly by a Swiss employer benefit from full free movement, outside the quota system. Third-country specialists remain capped; the 2026 national quotas stand at 4,500 B residence permits and 4,000 L short-term permits, so a hire from outside the EU and EFTA needs a longer lead time and a documented case.
Contractors cover the build phase, and permanent hires own the system afterwards
Contract or permanent is a decision about the life of the system. A contractor suits a defined build with an end date: a data platform migration, a first model in production, an audit of an existing pipeline. A permanent hire suits everything that follows, because a model in a regulated process needs an owner who monitors it, documents changes and answers the auditor. Edward Galle’s guide to build versus buy talent decisions sets out the wider trade-off.
Swiss law shapes how the contract route works. Under Article 12 of the Federal Act on Recruitment and Staff Leasing (AVG, Arbeitsvermittlungsgesetz), firms that lease workers to third parties on a commercial basis need a cantonal licence, and Article 39 sets fines for knowingly using an unlicensed provider. For individual freelancers, the compensation offices of the old-age and survivors’ insurance (AHV) decide self-employed status case by case for each activity, on the economic reality over the contract wording, and working for a single client is usually treated as employment. A long-running, full-time “freelance” data engineer embedded in one team is therefore an employment relationship in practice, and the firm should structure it as one.
A clean handover closes the build phase. In Edward Galle’s view, the contract for any external build in a regulated function should name the internal owner from day one, require the contractor to produce the documentation the quality unit will inspect, and reserve the final weeks for shadowing. The permanent hire then inherits a system with its records, and the employer avoids paying a contractor indefinitely to remember how it works.
Upskilling internal staff fills user roles faster than the external market
The jobs.ch survey found 33% of employees answering “yes, certainly” when asked whether they feel well equipped for AI, 42% already using AI as a helpful tool at work, and 57% training themselves, often informally through media, online content and trial of new tools. That informal learning is a pipeline employers can formalise. For AI user roles, the fastest source of talent is often the current team: a regulatory specialist who already writes structured submissions, or a brand manager who already runs CRM segments, can add AI skills inside a known process and keeps the domain knowledge an external hire would take months to acquire.
Builder roles follow a different logic: the external market supplies experienced data engineers and machine learning engineers, while internships and graduate programmes offer a junior pipeline for firms willing to train. A balanced plan hires a few builders externally, upskills users internally, and measures both groups against one skills framework that human resources (HR) and the business own together.
Employers who define the role first fill AI positions fastest
As of September 2026, the evidence points to one sequence: define the role as user or builder with its first deliverable, write the ad around that deliverable, screen on evidence, move inside the market’s window, pay senior specialists on an expert track, and treat the existing team as the first source of AI users.
Hiring managers can decide first whether the hire will use AI or build it, name the first deliverable of the first ninety days, and write the ad around that deliverable, with the regulated context in its opening lines.
HR and talent acquisition teams can replace title filters with an evidence screen of three to five capabilities, set one case study that mirrors controlled work, and run the whole process inside the window that the time-to-hire benchmarks for regulated roles describe.
Compensation leads can price senior data and AI specialists on an expert track that pays above middle management, as the SwissICT 2026 figures show, and extend the existing subject-matter expert ladder to data roles.
Candidats can bring one artefact to the interview, such as a repository, a validation summary or a short written case, that proves the capability the role needs on day one.
Edward Galle runs Brand Tech Specialist and Revenue Operations (RevOps) searches for Swiss pharma, MedTech, life sciences and technology employers, work where domain fit decides most hires, and supports HR teams through its HR consulting and workforce systems (HR IS) services. Specialists open to a move can submit a curriculum vitae to Edward Galle ou browse current openings.
This article has been reviewed by Haider Alleg, Managing Partner at Kainjoo SA.
References
- JobCloud. jobs.ch KI Report 2026. June 2026. https://www.jobcloud.ch/wp-content/uploads/2026/06/jobs.ch-KI-Report.pdf
- Federal Statistical Office (FSO). Zunahme der Beschäftigung im 2. Quartal 2026. 27 August 2026. https://www.admin.ch/de/newnsb/3iJDgJ638lidRZ0ahW4Vp
- Lucerne University of Applied Sciences and Arts (HSLU). Swiss AI Jobs Report 2025. 17 September 2025. https://www.hslu.ch/-/media/campus/common/files/dokumente/h/1-medienmitteilungen-und-news/2025/i/ai-jobs-report-2025-hslu.pdf?sc_lang=de-ch
- SwissICT. Saläre der ICT 2025: Moderate Lohnentwicklung und vertiefte Einblicke in Anstellungsbedingungen. 2 September 2025. https://www.swissict.ch/salaere-der-ict-2025/
- SwissICT. ICT-Salärstudie 2026: Post-Corona-Boom vorbei, Top-Spezialisten legen zu. 25 August 2026. https://www.swissict.ch/ict-salaerstudie-2026/
- Moneycab. Saläre der ICT 2025: Moderate Lohnentwicklung und vertiefte Einblicke in Anstellungsbedingungen (SwissICT salary study 2025). 2 September 2025. https://www.moneycab.com/it/salaere-der-ict-2025-moderate-lohnentwicklung-und-vertiefte-einblicke-in-anstellungsbedingungen/amp/
- Buholzer, D. KI-Spezialisten stellen die Lohnhierarchie auf den Kopf (SwissICT salary study 2026). finews.ch, 26 August 2026. https://www.finews.ch/news/finanzplatz/73353-ict-salaere-schweiz-ki-spezialisten-loehne-manager
- IT Reseller. Schweizer ICT-Saläre stagnieren. 26 August 2026. https://www.itreseller.ch/Artikel/106072/Schweizer_ICT-Salaere_stagnieren.html
- Canton of Zurich, Office for Economy and Labour. Kontingente 2026. 5 January 2026. https://www.zh.ch/de/wirtschaft-arbeit/erwerbstaetigkeit-auslaender/kontingente.html
- Swiss Confederation. Bundesgesetz über die Arbeitsvermittlung und den Personalverleih (AVG), SR 823.11, version of 1 July 2021. https://www.lexfind.ch/tolv/215326/de
- AHV/IV Information Centre. Merkblatt 2.02: Beiträge der Selbstständigerwerbenden an die AHV, die IV und die EO. November 2025 edition, status 1 January 2026. https://www.ahv-iv.ch/p/2.02.d