Digitale Transformation
„Interpretable AI has entered the mainstream.“
Everyday AI
Florian Douetteau, CEO of Dataiku appeals for data in the hands of many, NOT a few, if the use of AI is to become a real business asset!
Mr. Douetteau, what are the new trends driving digital transformation right now?
Surely that we now have many success stories of enterprise AI at scale – there is little doubt about the business value delivered!
This creates competitive pressure for the organizations that have not yet leveraged the opportunities. Across industries, we see that data-driven decision making can bring business to the next level. One example for this is GE Aviation where we enabled fast and better decision making in different functions within the company: engineering, supply chain and even finance.
And this happens not just there: Business functions found in all industries are systematically going beyond first proof-of-concept use cases on advanced analytics and are now talking about their success in production. This is also the point when more advanced questions like value generation reporting, lifecycle management, governance topics or extended scopes to the entire business come into the picture – and these are questions which may require board-level attention, as they go beyond a lone cowboy.
Florian Douetteau
What is the importance of democratizing data in the context of AI projects?
Without democratization of data you run the risk of having teams of data scientists in a silo or even their ivory tower. Doing AI projects like that is certainly one option… but you are missing a lot of potential in the organizations.
Besides that, the market of experts in the field is tight and we learn from many of our customers that simply hiring more data scientists does not scale.
So why not multiply the impact of AI value generation by involving all the business experts in the process? This means that you enable a larger set of people to deliver business value in shorter time frames. Basically you leverage the collective intelligence of the organization this way.
The expert teams are in no way obsolete – they play a vital role in enabling quality control, coaching people within the entire organization and paving the road for even more advanced solutions.
For example, Unilever expert data scientists built a social media sentiment analysis plugin for use by market-level analysts.
Without democratization of data you run the risk of having teams of data scientists in a silo or even their ivory tower.
What are the competitive advantages of companies that integrate all departments and employees into AI projects?
That is a great question! By actively involving your workforce in such an initiative you surely transform a diffuse “we want to be data driven” into concrete opportunity for all the subject matter experts. Once business functions are part of creating AI projects and giving them a transparent view on what is done also increases trust in the insights made visible by AI. By embedding this into a streamlined workflow you can also keep track of project management and governance across teams as well as compliance.
We believe that unleashing the collaboration is only possible when these diverse personas work together hands-on to deliver working AI. This leads to better design, faster deployment, and greater value at lower risk.
Etihad Airways has, for example, developed a forecasting model to optimize the staffing of check-in desks, which realized seven-digit savings and also decreased waiting times resulting in increased customer satisfaction.
What prerequisites must be created for this and how do you support your customers in this?
As Dataiku has evolved over time as a company, we have invested in going beyond pure software selling to include things like data integration, teams training (we’ve built a Data Academy to provide our users with scalable learning of our product Dataiku DSS) and business vision around topics such as data & AI governance.This also includes pre-built industry solution recipes that help to speed up the adoption of AI in the organization even more.
So we’ve developed expert teams all working towards the same goal of helping our clients achieve success.
Why is the integration of AI into all business processes important for the sustainable development of companies and their business models?
Clearly to enable future business success! There is a broad consensus that use of AI speeds up transformation of entire industries through cost reductions, better risk management and operational efficiency optimizations. This leads to growth and enables innovation which is underlined by success stories of our customers.
Ask yourself: What if your competition would successfully use AI and your business model is at stake? It is kind of the evil twin of the opportunity cost associated jumping on any industry trend that pops up.
Interestingly enough, companies choose Dataiku when they have this exact
hesitation but also realize how critical AI and data science are to propelling their business forward. In the next phase, scaling AI from an isolated, obvious use case towards the whole enterprise is hard, but it’s the reason Dataiku was started in the first place.
„We want to help our customers to create transparent, repeatable, and scalable AI and analytics programs – not just another PoC!“
What tasks does your AI platform perform in this context?
Our core mission is to make the use of data and AI accessible to everyone so that they can achieve exceptional business results – all while allowing for control, governance, efficiency, and speed.
We’ve found that many organizations spend a great deal of money on proof-of-concepts (PoC) that never make it into production. This is neither effective nor efficient. Experimentation is important, but one should keep cost in mind before this becomes unsustainable. If you just think about the cost of implementing a single use case, it does not capture the whole picture: Managing the life-cycle of a model, deploying it to production, monitoring its performance and associated machine-learning operations (MLOps) is where we can support with our systemized approach.
Because we appeal to no-code and low-code in addition to full code users, we’ve invested in our Dataiku Academy, tutorials, and self-service content, along with our Customer Success and Implementation teams to help our customers throughout their journey.
We want to help our customers to create transparent, repeatable, and scalable AI and analytics programs – not just another PoC!
About Florian Douetteau
Florian is the Chief Executive Officer of Dataiku, the platform democratizing access to data science and enabling enterprises to build their own path to Artificial Intelligence. The company has raised $145M to date and has spurred this journey to Enterprise AI for hundreds of large enterprise customers around the world.
Florian started working in the startup world at age 20 and hasn’t stopped since. He’s fond of creative writing (see AI Musings for a small taste), playing with real and artificial languages, as well as any board or screen games where he still has a chance to beat his kids. He’s also committed to Paris, a city where he was born, raised, loved and lives; Florian invests in and helps companies and tech founders as a small contribution to the growing Paris tech ecosystem.
What is your mission, here with regard to your guiding principle „Everyday AI, Extraordinary People“?
We believe leveraging AI at a large enough scale to become an organizational asset requires data democratization. That is, data in the hands of the many, not the elite few! This of course has to happen within the bounds of access restrictions and governance.
As such, we built a platform to enable people wherever and whoever they are (technical or non-coders, data scientists, engineers, architects, or analysts). Consequently, we provide a central location for distributed or remote teams, providing resources to work faster and smarter together for a more data-driven organization.
Times of economic change tend to expose companies that aren’t able to easily adapt, which is why agility and elasticity in AI efforts are critical, now more than ever.
What is your advice to companies and organizations that are just setting out to become „data-driven“?
Get started – and that means by actually implementing use cases!
Believe it or not, the customers we work best with are the ones that think they are not ready to do so and they need more time for planning, analysis or technical infrastructure. They see a diffuse goal of “data driven” being just technology focused and less of a people or change management topic.
What this means is that our bread and butter is identifying how and where our customers can make the most advances in AI aligned to their own strategic goals – of course respecting governance and compliance!
Dataiku is built to be embedded within your culture, demystifying AI and allowing everyone – regardless of coding ability – to leverage the power of data science to achieve business value.
You are one of the world’s most successful AI providers, what innovations can we expect in the future?
Our latest product innovations revolve around 3 strategic topics:
- Faster development and testing of robust and transparent models with advanced capabilities for the entire lifecycle
- Easier and even more scalable management of data projects beyond single players, teams and locations
- Further democratization of AI via augmented processors and accessibility to non-coders.
Dieser Beitrag steht unter Creative Commons CC BY ND zur Verfügung.
New Work
Als Mitarbeiter- bzw. Employee-Enablement wird die Praxis beschrieben, Mitarbeiter möglichst vielseitig zu schulen, um sie so im Arbeitsalltag polyvalenter zu machen. Enablement befähigt Mitarbeiter, fortgeschrittene Aufgaben auszuführen und komplexe Tasks zu meistern. Dadurch können Aufgaben im Laufe der Zeit immer flexibler delegiert werden, Kompetenz-Hierarchien verflachen und die Mitarbeiter bekommen das berechtigte Gefühl, dass in sie investiert wird, was die Wertschätzung ihrerseits dem Unternehmen gegenüber stärkt. Enablement sorgt für hohe Bindungsraten, da eine solche Unterstützung den Mitarbeitern das Gefühl gibt, dass ihre Leistung anerkannt wird.
Security Management
Digitale Unternehmenskultur
Cloud Technologien
Machine Learning / Deep Learning
Big Data
Artificial Intelligence / Küntliche Intelligenz
Data Science
IoT und IIoT
Mittelstand
Automatisierung / Robotik
Datenbasierte Geschäftsmodellinnovationen
Prof. Dr. Dr. Thomas Schildhauer
Verkehr / Mobility
Handel
Die digitale Transformation gewinnt immer mehr an Fahrt, dabei stehen momentan technologische Aspekte noch an erster Stelle. Betrachtet man jedoch die Innovationskraft von Unternehmen, so spielt die Unternehmenskultur eine tragende Rolle. Es wird deutlich, dass beide Themen eng miteinander verknüpft sind. Unternehmen, die nicht auf die digitale Unternehmenskultur setzen, können Innovationen nur schwer und sehr langsam umsetzen. Aber wie kann es Unternehmen und Führungskräften gelingen, den „Cultural Change“ erfolgreich zu gestalten? Also zunächst funktionieren digitale Arbeitsweisen nicht in starren Hierarchien und abgeschotteten Silos.
Finanzwesen
Bildung / Lernen 4.0
HR beginnt beim Menschen
Wachstum durch KI
CC BY-ND 4.0 DE
Sie dürfen:
- Teilen — das Material in jedwedem Format oder Medium vervielfältigen und weiterverbreiten
- Der Lizenzgeber kann diese Freiheiten nicht widerrufen solange Sie sich an die Lizenzbedingungen halten.
- Bitte berücksichtigen Sie, dass die im Beitrag enthaltenen Bild- und Mediendateien zusätzliche Urheberrechte enthalten.
Unter den folgenden Bedingungen:
- Namensnennung — Sie müssen angemessene Urheber- und Rechteangaben machen, einen Link zur Lizenz beifügen und angeben, ob Änderungen vorgenommen wurden. Diese Angaben dürfen in jeder angemessenen Art und Weise gemacht werden, allerdings nicht so, dass der Eindruck entsteht, der Lizenzgeber unterstütze gerade Sie oder Ihre Nutzung besonders.
- Keine Bearbeitungen — Wenn Sie das Material remixen, verändern oder darauf anderweitig direkt aufbauen, dürfen Sie die bearbeitete Fassung des Materials nicht verbreiten.
- Keine weiteren Einschränkungen — Sie dürfen keine zusätzlichen Klauseln oder technische Verfahren einsetzen, die anderen rechtlich irgendetwas untersagen, was die Lizenz erlaubt.
Freie Inhalte
Alle Beiträge stehen, sofern nicht anders gekennzeichnet, unter CC BY-SA 4.0 DE zur Verfügung. Sie dürfen diese vollständig unter Quellenangabe und Nennung des Autors sowie einen Link zur Textlizenz übernehmen.
