Discover this year's submissions to the Dataiku Frontrunner Awards and give kudos to your favorite use cases and success stories!READ MORE

AstraZeneca - Democratizing AI Through Upskilling and Self-Serving Citizen Data Science Tools


Callum Connolly, Data Product Consultant
Nick Brown, Head of AI & Data Science
Yaseen Ladak, Director - Enterprise AI Products & Technology
Hisham Jafar Ali, Associate Director - Enterprise Data & AI
Alice Smith, AI Platform Engineer
Kathiresan Anandhan, AI Support Engineer
Bhanuprathap Kari, Platform Engineer
Karthik Ramasamy, Scrum Master
Ritika Jain, Delivery Manager - AI and DS Platforms
James Harpin, Solution Architect
Curtis Scholey, Business Analyst
Haseeb Ahmed, Enterprise AI Platform Engineer

Country: United Kingdom

Organization: AstraZeneca

AstraZeneca is a global pharmaceutical company with a major UK presence. Our purpose is to push the boundaries of science to deliver life-changing medicines. The best way we can help patients is to be science-led and share this passion with the scientific, healthcare, and business communities of the UK.


Awards Categories:

  • Most Impactful Transformation Story
  • Excellence in Research


Business Challenge:

AstraZeneca’s IT2025 vision is that 80% of our core processes will be enabled by AI to completely transform how we deliver life-changing medicines to the public. However, with a limited number of AI experts in the AZIT organization, business units resorted to facing long waits before their models/pipelines could be built or paying hefty consultant fees to bring in outside consultants.

The solution to this challenge is the democratization of AI within AstraZeneca. By developing the skillsets of our people on the ground and making use of self-serving ‘citizen data science’ tools like Dataiku, we can enable our analysts, scientists, and statisticians to implement AI into their everyday work. That way, anyone can be the next person to bring AI to enable your day-to-day work, whether they are trying to forecast supply chain disruptions, use algorithms for drug discovery, or predict how many sales they’re going to make this summer. They are all part of the movement to democratize AI.

Taking our IT organization for example; typically a project requiring data science, ML or high level analytics on data collected by the business would be outsourced to the CFIT (customer facing IT). However, as there is a short supply of people who can carry out this work and an abundance of research and high-level projects, the time to value can be quite large and in some cases affect realizing new drug discoveries for patients. We see this throughout our operations groups including supply chain, inventory analytics, quality management, and procurement; each of which now have a place on our central instance running their own data science and analysis projects with the help of Dataiku placing the time to value back into their hands.


Business Solution:

We started the Dataiku initiative in early 2020 with a pilot group of 3 users within the AstraZeneca R&D organization to prove out the tool with a varied use case selection. The tool was received with glowing regard from the majority of internal users and was then rolled out to more diverse business areas.

Once we received the go-live mid-2020, our focus shifted towards configuring the tool to fit the needs of our users and to increase the visibility of the tool within the wider AZIT organization. We now boast a total of 200 users across the R&D, commercial, and Operations markets, with the majority of our users sitting in the commercial space where we facilitate the following markets: Global analytics, Australia, France, Russia, Switzerland, UK, Italy, and Spain. These 200 users contribute to a total of 50+ projects — all contributing to the shift in focus to start using AI in our everyday processes.

Time to value is a huge determining factor for the groups that come on platform, on of our EU/Canada leads Carlo Zanardi describes it so “It took minutes to connect it to the global data already available and it is a perfect fit to close a gap we had in the advanced analytics space”, “What is more relevant is how quickly we could start to use it (we do not need weeks of training)”

Another factor would be the drive for re-use in AZ, to not 're-invent the wheel', Hugo Cavalera a data scientist from French commercial testifies the following; “Beyond its ability to build Advanced Analytics pipelines, Dataiku helps us document and centralize data initiatives. It provides capabilities to track, comment, and describe the various artefacts created by the teams so that they can be reused across the organization”


Value Generated:


Since our go live in 2020, we have grown our community of citizen data scientists by nearly 200 active users creating data science and analytics projects in around 25 different business areas across AZ spanning from commercial international/European markets, Operations, Enabling units, and R&D with 50+ projects between them.

We have refined how we onboard and engage with these business units operating from a central team who handles front-end and back-end applications for automation and support, through a custom onboarding pack, onboarding portal on the AZ/Dataiku Academy, and AZ-specific documentation to enable the users as quickly as possible reducing the time to value significantly. “With any other tool I could use (known & available), this work would have taken me 5 times more time or even more” – Romain Dubois (Operational Excellence Business Partner - AZ).

Time to value is a huge driving factor in the use of Dataiku. With a low technical barrier of entry, we have found it easy, especially on the business side of the organisation, to enable our users and save in total over the last 2 years 1000 days of work based off user testimonials.


Value Brought by Dataiku:


As part of our initiative to enable 80% of our business process by AI and decrease the time to value for data science projects we keep constant engagement with our users to understand how their day to day analytics have changed through use of Dataiku. Below are a few key highlights from these time/money savings – all in all we have saved as an organisation over £1.5 million through switching to cloud computing and bringing the analysis in house through empowering our citizen data scientists.

Oncology Business (RWE)

  • Saving £20k per project (around 1/3 cost saving)
  • Potential saving of £200k

UK Commercial

  • Broadened the audience to 120+ users from R&I Field Force, Finance, and Marketing
  • Potential saving of £ 300k p.a.

Australia Commercial

  • Decreased implementation time from 126 days to 42 Saved £290k moving from the standard development process

Operations (Inventory Analytics)

  • DSS is helping to reduce file split and creation by 2-3 days manual work every budgeting cycle carried out by 2 person, with added benefit of accuracy and flexibility (24 days saved annually)

Operations Supply Chain

  • Dataiku is the only tool not requiring IT support to build a solution
  • Using Dataiku to analyse historic data and identify delays saving 50 hours per week across all CMO’s Global Commercial: Saving 20-40% of time spent by sales people driving to users, 1000s of man-hours saved monthly

France Commercial

  • A total of 60 days saved per annum across the team and their projects for a total saving of €54k per annum in resource fees

Spain Commercial

  • Increased our efficiency while allocating materials from 88.8% to 90.8% for our client Symbicort allowing them to better serve their demand and saving money on resourcing
Version history
Last update:
3 weeks ago
Updated by: