InsightsAdopting AI into organisations: carrying past lessons into the future
Share this article
Reflecting on projects and organisations I’ve worked with in the customer strategy and marketing technology space, there are new challenges daily, but how organisations harness data feels like Groundhog Day. From changes of funding models and head counts to the centralisation and decentralisation of teams, it feels like the value of the Data and & Analytics (D&A) functions is still in a phase of being scrutinised heavily, with some businesses unable to unlock the magic that was promised many years ago.
In 2024, AI is everywhere, creating both excitement and fear. The big question I often hear is: “How do we use AI to stay ahead of the curve?”. The lack of knowledge around the limitations has created a sense of infinite opportunity. The rate of change is rapid and big players are getting involved. It’s easy to see how organisations can feel like a deer in the headlights on what to do, afraid to be left behind.
The risk is empty promises, money wasted and no results. There are lessons to take from the D&A revolution to help guide us in the AI era, however. Some of the key themes from successful D&A transformations I’ve been part of at CACI that are relevant for AI adoption by analytics and data professionals in the customer space have been:
- Establishing value
- Data literacy
- Data modernisation
- Evolving the process
Establishing value
Establishing value is pivotal to getting the business along the journey by helping stakeholders understand how data and analytics helps their bottom line. Stakeholders don’t want a handful of numbers, they want the capability to make better decisions, execute more efficiently and deliver greater impact. Therefore, if a predictive model has been created to determine when is best to target a customer sale, the job is not done. The next steps are to substantiate what this can mean for the business in terms of opportunity, followed by activating it to drive and prove that the sale has happened.
This will be similar with AI. It is key to start by defining the use cases and business challenges to be addressed. Once this is understood, AI initiatives can have buy-in and be driven more easily. It doesn’t require a large roadmap, a series of proof points and steps to prove value is more than enough. Establishing what value is and demonstrating it unlocks the licence to move forwards with smaller, incremental steps.
Data literacy
Increasing data literacy is key to establishing a two-way conversation between stakeholders and D&A ambassadors. For stakeholders, it allows them to define their ambition and utilise D&A outputs to deliver to that ambition. For D&A ambassadors, it’s talking the language of the business, contextualising the day-to-day impact. Interestingly, working on this in the past has ended up with stakeholders mentioning “data” less.
For example, I worked with marketers who wanted to understand the opportunity in terms of how many customers they were going to reach. “What about the data?” was banded around, which can mean different things to different people. After helping educate them on the role of understanding counts and what that means for volumes, the language shifted to “volumes of unique eligible customers who will receive the campaign”. The less the conversation becomes about “data” and more what it means for customers while knowing the considerations with respect to data, the more effectively the business can reach its outcomes with less confusion and at a greater pace.
With AI, the role of ambassadors for data, analytics and AI is to always be translators, empowering users to understand and carry the conversation in all directions. That means fostering a culture where there is specific training for different stakeholders, tailoring how you talk to the stakeholder’s world and keeping at the forefront of developments to help people understand what AI means for both their day-to-day and future.
Data modernisation
I’ve often seen organisations leapfrogging with their technology capabilities or implementing data science models only to realise that the integrations were not set correctly and the data itself was not fit for purpose. There is the assumption of quality of data and that all tools are fit for purpose, however, data management and governance practices that have not evolved to meet requirements risk creating low quality data, which will affect outputs and create a lack of trust in the data and models. Furthermore, low data accessibility, exasperated by poor data management, can increase latency and make the speed to value slow and painful. These areas are typically not what stakeholders are thinking about and often results in large-scale data transformations becoming dead in the water.
Data modernisation requires reviewing infrastructure and governance so that processing and storage happens closer to where decision-making happens, improving speed, reducing cost and closing silos. Focusing on access, quality and efficiency will enable AI to be integrated in a way that is usable and scalable. Moreover, as AI application increases, AI-focused data management practices will significantly improve accuracy and performance of the models, which is crucial when productionising AI.
Although it may not be pretty or exciting for the end users, addressing data modernisation must be a key priority for D&A and AI ambassadors. There will be challenges in helping organisations understand the ramifications of substandard data management and governance practices. Tackling these issues head on will improve the time to value for AI and mitigate issues with quality, cost and output. Beyond this, modernisation of data governance must venture further– with a strict focus on ethics and compliance– by assuming the role of PII within the organisation and how that is used with AI, if using external technology.
Evolving the process
Once there is buy-in and a return on the D&A initiatives is recognised, interest and further investment will then be generated from the organisation. The next part, scaling, is sometimes the hardest step. In my experience, those who reach this point likely had smaller, autonomous teams tackling the D&A transformation. Moving forwards requires ongoing attention and adaptation, with the trend being to create specific roles and departments. While this can make sense, the risks include siloing teams and shifting focus from business outcomes to becoming more about delivering tasks.
The same will apply to AI, where it’s tempting to have an ‘AI department’. The balance that must be struck is the ability to deliver business outcomes versus the need to nurture specialisms to ensure that there is growth for the individuals, a combined view on the future and enforcement of best practice. This will emulate cyclical trends of centralisation and decentralisation of teams. This is not a bad thing– it’s okay to constantly evolve and adapt operating models around business needs. AI is unique in that it will become more pervasive in the day-to-day, so while AI technology may be centralised, its use will seep through the whole of an organisation.
How CACI can help?
Despite AI feeling like the next revolution, for some, it’s an evolution of data and analytics. We are in a period where D&A is being scrutinised in terms of its value, but the question is not being asked of AI just yet. It will be, and the themes above will gear you towards being able to drive ROI.
To learn more about how CACI can drive value with AI in driving value from your customers, contact us today.
Related posts:
Share this article