Skip to content
Marsanne Studio
← Back to home
← The AI watch

AI watch by industry

AI in software, IT services and telecoms

Adoption of AI in software, IT services and telecoms is rising fast but unevenly. In France, 18 % of firms with 10 or more employees used AI in 2025, up from 6 % two years earlier, but the rate reaches 58 % among large firms against far lower levels for smaller ones, according to Insee. Telecoms operators and software vendors are deploying agents at scale, such as Orange España’s more than 1,000 customised AI agents, yet a KPMG study points to persistent difficulties in scaling such projects across organisations, and most large French groups confine agentic AI to prototypes, according to the user association AUFO.

Adoption varies sharply by company size

National data confirm a widening gap by company size. Insee’s TIC survey found that 18 % of French firms with 10 or more employees used at least one AI technology in 2025, up from 10 % in 2024 and 6 % in 2023, but the rate reaches 58 % for firms of 250 employees or more, far above smaller firms. In Spain, 20.3 % of firms with 10 or more employees used AI in 2025, according to Eurostat data cited at the AmChamSpain Global Trends 2026 conference, while across the EU adoption ranges from 55 % among large firms to 17 % among SMEs of 10 to 49 employees. In France, only 15 % of small firms report structured AI integration, against 58 % for large groups, according to an Ipsos BVA survey.

Our summaries: Insee, Demócrata, Solutions Numériques

Mid-size firms report concrete, measured gains

Several mid-size firms report measurable results. The French online accounting firm Dougs, which says it serves more than 50,000 businesses and reached €50 million in annual recurring revenue in 2025 (growth of 50 %), raised €50 million partly to expand AI-driven processing of accounting files. The Spanish IT distributor PcComponentes, with sales of €825 million in 2025, now sells servers letting firms run AI models locally without sending data to the cloud. The consultancy Zenital describes a client project where automation cut a task from 240 to 80 hours. Red Points, grown from 10 to more than 300 staff, runs 90 million fraud detections a year using AI it has built since 2015. Galeo, serving factories and hospitals, grew its own headcount from 39 to over 60 in twelve months.

Our summaries: FrenchWeb, Xataka, Revista Byte TI, Emprendedores, Computerworld España

Telecoms operators deploy agents at scale, but scaling stays hard

Telecoms operators illustrate both the scale of deployment and its difficulty. Orange España has expanded its alliance with Google Cloud to put more than 1,000 customised AI agents into operation, aiming to industrialise its AI processes with Gemini Enterprise. Yet a KPMG study found that telecoms operators generally are accelerating AI adoption while struggling to scale it across their organisations. A study by IFS and The Futurum Group found demand for specialised AI reaching 69.5 % among telecoms respondents, among the highest of any sector surveyed, close behind construction and engineering at 72.4 %.

Our summaries: Computing, elEconomista, El Ecosistema Startup, IT User

Large groups limit agentic AI to prototypes; most firms stay cautious

Even among large organisations, agentic AI remains limited in practice. The co-presidents of AUFO, the French association of Oracle users (1,000 members, 150 companies), say agentic AI is not yet deployed at scale among French clients: large groups keep it to advanced prototypes, mid-size firms are only discovering the topic, and human oversight is maintained throughout. In Spain, the Banco de España found that 60 % of firms already using AI remain in a testing phase, with only 6 % reporting relevant use. Denodo co-founder Alberto Pan argues that scaling AI beyond isolated use cases requires a new data architecture. In France, 71 % of non-adopting firms cite a lack of perceived usefulness and 54 % cite a lack of expertise, according to Insee.

Our summaries: LeMagIT, Demócrata, Computerworld España, Insee

Security and governance of AI agents need active management

Deploying agents raises sector-specific governance questions. A Cloud Security Alliance study cited by security vendor Semperis found that nearly two-thirds of organisations have suffered a security incident linked to AI agents, 92 % have staff who installed AI on local machines with access to sensitive keys, and 82 % host unmanaged agents. The French group Spie, testing Sekoia’s agentic security operations centre since July 2026, reported 96 % of around 10,000 verdicts left unchanged by analysts in September, with Sekoia citing an average investigation time of 88 seconds across its base of 1,000 clients, though only 2 of 411 predicted true positives were confirmed. For smaller firms without an IT department, the Model Context Protocol promises to cut integration costs between AI and existing management software.

Our summaries: Solutions Numériques, Silicon.fr, IT for Business, Journal du Net

Articles cited

Another question?

Claude searches the articles in the watch and answers you, citing its sources.

Answer written by an AI model (Claude, by Anthropic) from the articles in the watch only: check the figures in the sources. Questions are recorded;learn more.

Method

Text written on 20 October 2026 by an AI model (Claude Sonnet 5, by Anthropic) from 80 AI watch articles on this theme, published between 8 October 2026 and 19 October 2026. Each paragraph links to the entries it draws on, and every number in the text was found in one of these articles before publication. The text is rewritten every week.

Other industries

Other topics