{"id":35871,"date":"2026-05-26T14:28:00","date_gmt":"2026-05-26T12:28:00","guid":{"rendered":"https:\/\/askme.it\/insights\/ai-coaching-for-pharma-sales-reps-what-is-really-allowed-in-italy\/"},"modified":"2026-05-28T17:54:54","modified_gmt":"2026-05-28T15:54:54","slug":"ai-coaching-for-pharma-sales-reps-what-is-really-allowed-in-italy","status":"publish","type":"insights","link":"https:\/\/askme.it\/en\/insights\/ai-coaching-for-pharma-sales-reps-what-is-really-allowed-in-italy\/","title":{"rendered":"AI coaching for pharma sales reps: what is really allowed in Italy"},"content":{"rendered":"
\n

Between July 2025 and March 2026, AI coaching platforms for the pharmaceutical sales force entered the enterprise production phase. Quantified updated its roleplay avatars in July 2025 declaring an ability to recognize and react to the detail aids shown during the simulation, with a claim of an average performance improvement of 20% and a reduction in onboarding times of 42% for new hires during a launch. Retorio declares a measured improvement of 16% after a single repetition session, on a sample of more than three hundred sessions in pharma clients including Merck and Amgen. Veeva brought AI Agents for Vault CRM to general availability on December 3, 2025, including Pre-call Agent, Voice Agent and Free Text Agent, with the promise of closing the observed-suggested-practiced loop.<\/p>\n

For Italian pharmaceutical companies the value leap is real, but the design of the deployment is less free than in the United States. Article 4 of the Workers’ Statute, in the text modified by the Jobs Act of 2015, requires that any system for recording or analyzing the worker’s conversation during work activity be the subject of a preventive union agreement with RSU or RSA, or in the absence of an authorization from the Labor Inspectorate. The findings of the Italian Data Protection Authority (Garante) in 2024 and 2025 have clarified that a “quality recording” system applied to the sales force falls within the perimeter of Article 4 even when the company presents it as a training tool. The deployment of AI coaching in Italy starts with union work, not with technology procurement.<\/p>\n<\/section>\n

\n

What AI coaching really does today<\/h2>\n

Pharma AI coaching platforms operate on three distinct phases of the visit cycle. The pre-visit phase is the domain of simulated roleplay: Quantified, Retorio and ZS’s conversational simulation modules allow the representative to rehearse their pitch in front of an avatar that recognizes detail aid content, responds with the conversational profile of a simulated physician and produces structured feedback on content, pace, objection handling and non-verbal posture. It is the most mature segment and the one that poses the fewest Italian regulatory issues: no recording of the real worker, no HCP data involved, no interaction with the company’s production system.<\/p>\n

The in-visit and post-visit phase is where the promised value grows and where Italian constraints make themselves felt. Veeva Voice Agent allows dictation of the call report immediately after the visit with AI processing of notes; Free Text Agent performs a real-time compliance check before saving. Conversation intelligence applied to real recordings \u2014 the promise that US platforms repeat in their claims \u2014 produces managerial dashboards with metrics on engagement, quality of objections handled, physician sentiment, alignment to the key message. This last component, in Italy, requires a union agreement. Without it, it is a legal exposure for the employer.<\/p>\n<\/section>\n

\n

Documented international cases<\/h2>\n

The most publicly documented deployments are in the United States. AstraZeneca put a machine learning system on field-coaching forms into production starting in 2019 and scaled it through 2024 across four thousand reps and six hundred managers, with a managerial function that allows observation of coaching patterns across the entire organization. Novartis used Quantified for mass onboarding during a launch, cited by the vendor’s materials as a reference customer. Merck and Amgen appear as Retorio customers. Pfizer has invested in AI simulation capabilities in its evolution of the commercial field force.<\/p>\n

None of these case studies is easily transportable to Italy without structural reconfiguration. The distinction between real visit recording \u2014 which in Italy requires a union agreement and the physician’s explicit GDPR consent \u2014 and pre-visit simulated roleplay is the first question to resolve before even evaluating the platform. The second is the physician’s consent to the profiling of their responses: the Garante has in recent years issued rulings in which it has qualified the profiling of HCP prescribing attitudes as an activity that requires explicit consent, not falling within the legitimate interest of pharma marketing.<\/p>\n<\/section>\n

\n

The viable path: roleplay and self-reflection without tracking<\/h2>\n

The Italian deployment of AI coaching compatible with the regulatory framework in 2026 is concretely designable. The first level is pre-visit roleplay on simulated avatars: no real data of the worker or physician involved, no Article 4 exposure, immediate deployment. The second level is the Pre-call Agent in Veeva style, which reconstructs the HCP context by drawing on CRM data and public external sources, with the representative always remaining the passive recipient of a suggestion, not the object of measurement. The third level is voluntary post-visit self-reflection: the representative autonomously records their own voice transcription of their own notes, on a system that does not coincide with the production system, and submits it to an AI agent that returns structured feedback on the content of their own conversation. The recording is by the worker on themselves, not by the employer on the worker.<\/p>\n

On top of these three levels, anonymized aggregate dashboards useful to the training manager can be built, but not individual profiling of the single sales rep. It is a tighter perimeter than what Quantified and Retorio sell in their marketing materials, but it is what can be defended in front of a labor inspection or a Garante audit. The performance value promised by AI coaching exists also within these boundaries: the difference between a representative who has rehearsed the pitch ten times on an avatar and one who tries it for the first time in a visit is measurable, regardless of how sophisticated the subsequent monitoring system is.<\/p>\n<\/section>\n

\n

The constraint that weighs more than the platform: adoption<\/h2>\n

The 2025 ISF Observatory, third edition published in June by Informatori.it and Bhave, photographs an Italian market in which 50% of representatives reach the goal of nine visits per day, the CRM adoption rate is at 77%, the average age is around fifty years, 69% have more than ten years of experience, those under 34 represent only 8% of the population, 86% participate in company training. It is a sales force that is mature, experienced, with a high baseline of professional autonomy and significant resistance to top-down monitoring. AI coaching platforms work in contexts where the representative perceives the tool as an amplifier of their own practice, not as a source of automated managerial judgment.<\/p>\n

The success declared by Aktana in its own publications and the one documented by AstraZeneca pass through an adoption threshold between 60 and 80%. Below that threshold the system produces alert fatigue, above it produces real performance lift. For Italian companies the operational constraint of AI coaching in 2026 is neither the available technology nor the cost of the license: it is the design of the introduction process, the dialogue with union representatives, transparency with the sales force, alignment of the incentive model. The +14% in sales of Aktana and the +20% performance of Quantified exist. They are not included in the subscription.<\/p>\n<\/section>\n","protected":false},"excerpt":{"rendered":"

Quantified declares a 20% performance improvement, Retorio 16%, Veeva moved its AI Agents to production in December 2025. The Italian limit is not the technology, it is Article 4 of the Workers’ Statute. What can be deployed without a union agreement and what cannot.<\/p>\n","protected":false},"featured_media":35873,"menu_order":0,"template":"","insights_category":[968],"insights_tags":[973,649,907,971,974],"class_list":["post-35871","insights","type-insights","status-publish","has-post-thumbnail","hentry","insights_category-ai-and-pharma","insights_tags-ai-coaching","insights_tags-change-management-en","insights_tags-conversation-intelligence","insights_tags-medical-representatives","insights_tags-workers-statute"],"acf":[],"_links":{"self":[{"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights\/35871","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights"}],"about":[{"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/types\/insights"}],"version-history":[{"count":1,"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights\/35871\/revisions"}],"predecessor-version":[{"id":35872,"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights\/35871\/revisions\/35872"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/media\/35873"}],"wp:attachment":[{"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/media?parent=35871"}],"wp:term":[{"taxonomy":"insights_category","embeddable":true,"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights_category?post=35871"},{"taxonomy":"insights_tags","embeddable":true,"href":"https:\/\/askme.it\/en\/wp-json\/wp\/v2\/insights_tags?post=35871"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}