AstraZeneca, the biopharmaceutical group headquartered in Cambridge in the United Kingdom, has certified more than 17,000 of its employees in AI competencies, and mandates that all staff above a certain grade reach at least Silver. That is how Fortune reported it on 30 April 2026, citing what the company’s Chief Financial Officer, Aradhana Sarin, told the magazine. If the question is how AstraZeneca certifies its staff in the use of artificial intelligence, the one-sentence answer is this: through an enterprise-wide upskilling programme, launched in 2024 and described by the company itself in April 2025, that certifies employees on a tiered scale running from Bronze to Diamond, so that each person holds a named level.
The number matters less than the design. Most organisations that have run an awareness campaign on new technology can say how many people attended. Far fewer can say what level each person reached, and fewer still have made a level the condition of a grade. AstraZeneca has done both: the first it has described in its own words, the second its Chief Financial Officer has described to Fortune.
Not who attended, but what level each named person reached. That is the difference a tiered certificate makes.
What AstraZeneca built, in its own account
In April 2025 three AstraZeneca executives, Cindy Hoots, Brian Dummann and Maciej Szymaszek, published a description of the programme on the company’s media centre. Launched in 2024, it gives all employees, “regardless of their experience or comfort level with technology”, the opportunity to learn how generative AI can be used in their everyday lives, and how to use it in a responsible and ethical way. The company says it partnered closely with Global IT, HR, Compliance and Business Functions to structure what it calls an enterprise-wide AI upskilling and literacy programme.
The format mixes keynote lectures and lab sessions with reflection activities and strategic thinking workshops. The digital and AI learning pathways are available to every employee in 12 languages, and the company says it is adding upskilling opportunities in local languages this year. Learners progress through what the company calls a tiered competency scale: Bronze, Silver, Gold, Platinum and Diamond, with the top levels described as advanced levels of accreditation introduced in 2025. By April 2025, AstraZeneca said, “Over 12,000 employees have participated to date”, earning Bronze, Silver and Gold certifications. A year later, Fortune reported that more than 17,000 employees had been certified in AI competencies. Fortune describes the framework as Bronze, Silver and Gold; the company’s own 2025 article sets out the five levels above.
One of the programme sponsors, and a co-author of the April 2025 article, is Cindy Hoots, SVP, Chief Digital Officer and CIO. Her aim is direct: “By upskilling our employees in AI at speed and scale, we’re future-proofing our workforce to make agile, informed decisions that drive both business and scientific breakthroughs.”
Why a tiered certificate is different from a completion record
A completion record says that a person opened a course and reached the end. A tiered certificate says that a person demonstrated a defined level of competence, and it says which level. The difference is small on a slide and large in practice.
For a head of learning and development, a tiered scale gives three things a completion report cannot. First, a way to differentiate: a person who builds prompts and checks outputs for a living can be held to a different level from a colleague who needs to recognise a risky use when she sees one. Second, a way to show progression: a Bronze holder in 2024 who reaches Gold in 2026 has a visible path, and so does the manager who approved the time. Third, a way to set a standard: once a level exists, it can become a condition of a grade, which is what Fortune reports AstraZeneca has done with Silver.
What the minimum level does for compliance
Pharmaceutical companies work under close regulation wherever they operate, and AstraZeneca notes that it was one of the first pharmaceutical companies to publish principles for AI and Data Ethics. Compliance is one of the four partners AstraZeneca names in structuring the programme, alongside Global IT, HR and Business Functions. That is the detail a head of compliance at any regulated employer will notice.
For any employer, a minimum level above a given grade turns a policy into a measurable state. When a question is asked about who in a function is trained to use these tools responsibly, the answer can be a list of named people and the level each holds, not an attendance count. And because the requirement is set by grade, it reaches leaders, not only new joiners.
The CFO’s own reading, as quoted by Fortune, is about adoption rather than enforcement: “I think people are really embracing AI and learning and developing their own skills.” The same report puts the number of active AI pilots across the company at about 1,000, and quotes Sarin saying that “the value capture happens when you convert those pilots into production and really embed that in the workflow”. A workforce certified to known levels is a reasonable place for that conversion to start.
What a head of learning at a regulated employer in Ireland or the UK can take from this
Few employers have AstraZeneca’s scale, and none needs that scale to borrow the design. Three choices travel well to a professional services firm, an insurer, a utility or a certification body of a few hundred or a few thousand staff.
The first is to define levels before building content. Bronze, Silver and Gold are only useful if each one names what a person at that level can do. Write those statements first and the curriculum follows.
The second is to make the top tier about judgement, not features. The tools will change several times before the programme is reviewed. The ability to recognise what must not go into a tool, to check an output before relying on it, and to know when to stop and ask, does not change with the release notes.
The third is to decide, with Compliance in the room, which grades carry a minimum. A minimum that applies to everyone is easy to announce and hard to evidence. A minimum that applies above a grade takes more explaining, but it is easy to evidence, because the population is known and the record is per person.
Where LearnFrame’s AI for Professional Practice programme fits
At LearnFrame, Dublin, Ireland, this is the territory of the AI for Professional Practice programme. Its scenario-based modules put a professional in front of the decisions that arise with these tools in regulated work, and record each decision and its reasoning, person by person. The output answers the question a tiered certificate answers: not who attended, but what each named person did and why. A sample of that record, for a fictional organisation, is at learnframe.com/ai-in-practice-record.
AstraZeneca’s programme is enterprise-wide, has run since 2024, and belongs to one of the UK’s largest companies. The lesson for everyone else is the shape of it: levels defined, Compliance among the partners, a minimum set by grade, and a certificate that names the person and the level. Those four choices are available to a learning team of any size.
The AI in practice policy record is what an organisation holds after every member of staff has taken the twenty-minute Module 1 exercise: executive summary first, the five decisions and the policy lines second, person by person. A two-page sample for a fictional organisation is on the site.
Sources: AstraZeneca, “How we’re upskilling and preparing our workforce to thrive in the age of AI”, by Cindy Hoots, Brian Dummann and Maciej Szymaszek, published 11 April 2025 (astrazeneca.com): the 2024 launch, the partners, the formats, the 12 languages, the five levels, the 12,000 participants and the Hoots quotation. Fortune, “How AstraZeneca’s 17,000 AI-certified employees are helping it reach a ‘stretch goal’ of $80 billion in revenue”, by Sheryl Estrada, 30 April 2026 (fortune.com), a secondary source reporting what the Chief Financial Officer told the magazine: the 17,000 figure, the Silver minimum above a grade, the 1,000 pilots and both Sarin quotations. Fortune describes the framework as Bronze, Silver and Gold. Both retrieved 2 November 2026.
For another employer that defined the skills before building the content, see PwC (2026): 30 skills for an AI-native workforce, and a rule that AI is never taught on its own. For a programme that reached every member of staff, see Santander (2026): AI tools for all 185,000 staff, mandatory AI training, and a €1 billion target. For how a professional’s judgement can be scored, see One test, five ways of scoring it. See more insights from LearnFrame.