How to gain more business impact from your data & AI investments? Start from the people!

Minna Kärhä Business Lead, Data Driven Business , Solita

Published 22 Feb 2024

Reading time 8 min

Do you feel your organisation’s investments in data & AI aren’t delivering appropriate value for your business? Or have your data & AI budgets been cut, making you as a D&A leader feel hopeless? We have discovered two typical challenges that organisations often face nowadays and three solutions to help fix these problems. In this blog, I’m introducing our practically established insights.

Many organisations currently face similar challenges related to how the different functions view and approach the topic of data & AI. First of all, the data management roadmaps and portfolio, owned typically by IT, tend to focus only on the technical enablers. These roadmaps and portfolios consist of projects that focus on the technical foundations including typically themes like cloud data platform, master data management, data quality and -governance, data cataloging and metadata management, and data literacy, which typically seems to translate into providing everyone with BI tool and some online training to get started. As themes, these are very relevant. But the challenge is that the concrete activities are focusing on which tools to choose and which frameworks to use. Nothing that really speaks to the businesses that are eventually funding these. The only one that business users typically understand is the BI licenses. However, the challenge is that they aren’t able to use the tools if they lack a proper understanding of the organisation’s language of data.

Next to these common technical enablers, every business function typically has a project portfolio as well. These consist of projects that are critical for business success but aren’t connected to those technical enablers in the IT’s portfolio. In addition to their development projects, businesses also have various ad hoc support requirements, linked to critical everyday business activities. Those needs include developing new insights from data utilising analytics and AI, getting access to new data that has been recently acquired or produced, fixing data issues that affect the ability to use the currently available data for decision-making and steering the business, and solving data related questions to help various type of business users to gain benefit from the data.

Do you find similarities to your organisation from these? So, what is the problem here, having a plan for IT and a plan for business?

The biggest challenge is that both scenarios compete for the same resources, skilled and knowledgeable people, time, and money. The paramount problem from the IT perspective is that they acknowledge the continuously growing dept that is making use of data labour some and slow. At the same time, the most prominent problem for businesses is the inability to run their daily activities without the appropriate data. Highly impacting their ability to meet their business goals.

Fundamentally, the most significant problem is that there is no common problem to be solved! IT’s data vision is “making data available and useful for all” while business vision says nothing about the opportunities to be utilised, provided by data and AI.

Neither IT nor business can solve their own problem alone, they both need collaborative actions with each other.

Bringing data consumers and data management together is hard, but possible!

Bringing people together is hard if there is no clarity on what’s in it for them. The data consumers in business aren’t interested in where their data is stored and through which tools it is processed before they get to use it. They want to get access to data that is relevant to them and helps them to do their daily work smoothly and better. They know that data already exists somewhere. (Otherwise, they wouldn’t know to ask for it). Their needs are met only with the ability to connect that data to the tasks they have at hand.

When the data consumers are frustrated, it most often relates to the slowness of getting the data they need to use. It can feel like watching a video getting loaded with an extremely poor internet connection. The viewers’ patience is tested when streaming content constantly pauses to load. You can almost see the full screen but not quite to be able to enjoy the video.


At the same time, without these fundamental data management capabilities in place, the work of data experts in IT will continue to focus on chasing after the repeating data issues, always running to extinguish new and old fires. When IT’s data experts aren’t able to fix issues with more robust solutions, implement the kind of data management tooling and frameworks they would prefer, they get frustrated. Feels like trying to run a marathon with high heels while the business is watching and yelling why are you so slow? It’s not just slow, but extremely dangerous for your ankles and hurts your feet enormously.

How to bring these two together, so that your organisation has enough focus and makes wise investment decisions to build a robust foundation for your organisation while at the same time enabling the daily business activities to succeed with adequate data? In other words, how to co-create a data and AI strategy that is deeply linked to the business objectives.

We believe that starting small and showing through example what good can look like is the most impactful way forward. Focus on what already exists and help people in the organisation to utilise and get value from that. It won’t only enable your organisation to get value from the investments already made but also provide the organisation with valuable learning opportunities for collaboration and using these existing capabilities to do their work more successfully. As the understanding in the organisation grows, there is likely higher interest in further developing the fundamental capabilities, including renewing some of the tools as well. Working together and learning together by having common objectives!


Three practical tips to get started

TIP1: Start from what you already have

You already have data and information available for several users and you have tools that enable these to be accessed and interpreted. Make sure that what you already have in place is utilised at the full potential:

Are people really making decisions based on the information they already have? Or are they just asking for more because they don’t know how to use the existing data? Support them to grow data literacy instead of becoming a “report factory”. It is not a rare case that there are 3 times more reports in the organisation than there are users.

Do people know what information is available to them? Help them with making information discoverable. And support the collaboration in knowledge exchange to turn the distributed information into a shared asset. By bringing people with different knowledge together, you can reduce the need to create new reports to answer the same questions multiple times. Time is saved for more value-adding tasks instead.

Educate the data consumers to analyse and refine the data and information themselves. Provide peer support and communities where learning to use data is comfortable and enabled. This doesn’t mean you give them the user credentials and access to the BI tool. In practice, you should find out the key users, who can eventually help others in the organisation, and guide them to become more self-sufficient in terms of using available data. Sit together with them, show, tell and gradually let them try by supporting them. The time used here will be saved multiple times later.

TIP2: Start by making data care a shared responsibility, integrated into other daily tasks

The cost of fixing poor-quality data grows, the further the data is in the value chain. On average, the cost of fixing the poor data quality after the data was produced costs 10 times more compared to ensuring the quality is sufficient when it is first produced. If poor-quality data flows into use, the cost is 100 times more. Instead of trying to work as data janitors by cleaning data after data producers have messed it up, IT’s data experts should start to build awareness and co-create with the data producers’ practices that help to keep the data quality at the needed level. Again, don’t start with tools or theoretical frameworks and tens of pages of documentation. Focus on the data quality and trustworthiness from the business perspective.

Facilitate collaboration with the data consumers and data producers to understand where the quality problems occur that have the biggest impact on users and business. Usually, you can start approaching the juiciest problem by listening to what people mostly talk about. Do you hear comments like “this is again messed up” or “we don’t use it because it can’t be trusted”? Who is saying these and on what occasions? Sit down with these people together and dive deeper into the source of the problem: What is causing those issues? Don’t just fix it when you first see it but trace it to the very source. When you have pinpointed the root cause of the problem, focus the resources on fixing that at the very start of the data value chain! Co-create solutions that help to keep the quality at a sufficient level together with the people producing the data in the source of the issue. You will likely not need to start with a new tool. Instead, people need to agree on how they will work to prevent the issues from occurring.

TIP3: Do a sanity check on your current data & AI architecture

Before planning on any new investments in technology, first assess your current tools. Are your current tools the right fit for their purpose in your organisation at the moment? If they aren’t used or providing value, why is that? A very common problem is that some of the tools been acquired to the organisation, end up not being used. The reason can be that they aren’t “liked”, simply because the purpose of them isn’t understood or people just can’t use them. In this case, just replacing a tool with another one won’t solve the problem. The new tool won’t be used either. To prevent falling into the same trap again, investigate what the real challenge is. Why don’t the intended users like the tool? Why don’t they use it? Find out ways to turn it around, answering questions: how could the tool help them to do their work better? Is there a need in the organisation for the functionalities this tool provides? Is the too complex to use? Why is that? Maybe someone is even trying to use the tool for the wrong purpose and therefore gets frustrated. Most often, people don’t need state-of-the-art tools. They need easy-to-use tools that help them forward with other tasks they have at hand.

Be courageous and instead of thinking and looking for a tool like others do start from the people! Because they are the most critical part of this. We have experience in guiding and coaching teams and organisations to discover where change is needed and how to integrate the desired ways of working into their daily tasks.

Read also what we did with Pfizer: Trusting the design process helped Pfizer create a user friendly data catalog.

Interested for more? Join our Crash Course on Data and AI strategy and hear Minna’s insights on the topic.

  1. Business
  2. Data