• Blog
  • Generative AI: what’s at stake for your business?

Generative AI: what’s at stake for your business?
Understanding the impact and potential value of AI

Jim Bretschneider
30 aug, 2023

Generative AI will make our professional lives easier, right? It offers us tools to enhance productivity, automate and enrich reporting, customer service, just to name a few. Or… are the risks involved bigger than we think? And how do we actually board that train and not fall off?

Jim Bretschneider is Executive Vice President Solutions at HSO and has been engaging with many customers’ executive leaders on the topic of AI. “The potential of AI is huge, but it’s not always easy to benefit and get the actual value out of it. We help bring executive teams to a certain level of knowledge about AI and in that process we address many questions and concerns.”

We asked Jim on what he sees as best entry points for companies, what he encounters as potential hurdles on the AI journey and how to overcome these.

“Your biggest concern should not be whether you can keep up with AI, but if your application and data platform is actually capable of supporting all innovation Microsoft is offering.”

Get your AI-briefing

Q: What topics come up most in these conversations with executive management teams?

A: What I notice is that companies are looking for a reliable partner that not only has the knowledge, but also understands what is needed to make AI work and scale in their industry and market. The majority of companies are at least experimenting with AI or already applying it in one or more of their working areas. However, they struggle with questions like “What are the risks of generative AI?” and “How can I address the concerns of my CSO?” But also questions like: “We have experimented with Copilot or chatbots, but how can we scale it?”

Q: When we look at the Microsoft AI suite, how can companies get this to actually work, safe and secure?

A: “First of all, the Azure AI offerings are pretty amazing and it’s developing so fast. Something we know from a solution preview today, might be live tomorrow. And yes, we see various interesting starting points. To address the topic of security and trust, in many companies, individuals have just started to experiment with OpenAI and they might find it useful. However, yes, your CISO is right, there are security risks, when chat history includes customer names or other sensitive information if this is not implemented the right way. Switching to Azure OpenAi or Bing Chat Enterprise, with commercial data protection are often the initial steps in the generative AI journey.

What many companies are obviously interested in is using AI to automate and improve Customer Service, chat with their data or solve for a soecific labor intensive usecase.However, all these entry points and possibilities are changing so quickly! I compare AI time with time in dog years. What used to take months and years when it came to developing business applications, just takes days and weeks for AI.”

Customer success with

These customers work with HSO on their Data and AI journey

Q: When it comes to coding, how will AI impact our customers?

A: “The impact for development teams is huge, within HSO as well as at our customers. Over 40% of code written today is generated by AI, but that is not the only change. Traditionally, teams started a project with whiteboarding, designing the process they wanted to automate and then eventually they tried to build the code and the algorithms. With AI, you start the other way around and validate if your AI model is able to provide the response you are looking for. When AI can generate that for you, it’s relatively easy to build the process around it. This will increase development capabilities and speed a lot.

Another advantage is the AI capability of documenting code, which often causes headaches – people are not following documentation guidelines or the developer of that specific piece of code has left the company. Today, generative AI can document the code from a technical perspective and often explain the use case."

Q: All these entry points seem pretty safe and relatively easy to start with. But what about the investment required?

A: "Many customers start experimenting with AI and Copilot themselves, without earmarking significant budget. But then they find out it’s hard, and the experiments do not deliver the expected results or ROI. That’s where I see our role as an experienced advisor, we have the tech and industry knowledge to make solutions ready to scale and deliver value. But this can’t be done without initial investment.

Of course, there are uncertainties and risks involved when a company starts investing in AI. We often don’t know the outcome of an AI project, since the developments go so fast. Therefore, in order to prepare your company for all that is going to be out-of-the-box available very soon, you may need to invest in data projects first. We also foresee that companies need to re-assess their use of ERP and CRM. If you have drifted too far away from the standards or are not using the applications as they were intended, Copilot will not work. Similarly, when you start chatting with your data, if your document archiving is not done properly, AI won’t understand it."

Collaboration is key

"From an organization perspective, there is a lot more collaboration needed between Business, Technology and Data & Security. Creating a virtual or formal innovation team with representation from all these areas is often a good approach."

Four people working around a table collaborating on a business value assessment

Q: Circling back to the first question, what makes AI hard?

A: “Research shows that a high percentage of AI and Machine Learning projects have been failing. The main reason for this used to be that organizations did not have the right data to train the model and thereby cause unreliable answers or unintended bias. In 2023 and beyond, moving to generative AI, reduces the dependencies on having the right quality and quantity of data to train ML models, but a new set of challenges emerge. If you want to implement Sales Copilot but you are not using Dynamics CRM as it was intended, or your document library is a mess, it will be hard to realize the full value.

My main takeaway is that, unless you're a technology company, you shouldn't worry about whether you can keep up with AI technology. Your main concern should be whether your organization, applications and data platform is actually capable of supporting all the innovation Microsoft is offering you. And then, the potential impact and value that AI can offer is amazing.”

Resources & Solutions

Read more about Data, AI and the Microsoft Azure platform

Video: getting started with generative AI using Azure OpenAI Service

One of the many videos from the Microsoft Build 2023 event.

Learn more

Contact us

Our experts are ready to help. Why not start with an executive AI briefing?

By using this form you agree to the storage and processing of the data you provide, as indicated in our privacy policy. You can unsubscribe from sent messages at any time. Please review our privacy policy for more information on how to unsubscribe, our privacy practices and how we are committed to protecting and respecting your privacy.