Article

What successful AI adoption might look like

Name: Implement Consulting Group
Price range: $

Unlocking the broader potential of generative AI

Published

26 November 2024

You might recently have acquired generative AI tools for your organisation, or maybe you are already exploring how this new technology can help boost productivity and quality in the short term and help inspire your employees to think about new innovative products and services in the longer term.

The potential impact of generative AI on global economic growth is estimated to be nothing short of huge^1,2 and organisations are picking up on the hype and have been scrambling to jump aboard the generative AI train; the percentage of organisations utilising generative AI in at least one business function has surged from 33% in 2023 to 65% in 2024³.

In parallel with the emergence and diffusion of generative AI into work at large, we observe an increasing number of organisations facing new challenges – these are challenges in securing broad buy-in and adoption of generative AI tools among the organisations’ employees. The reasons are technical, organisational and human, and that combination poses a unique challenge – but it is what we should expect from unique tools that have the potential to facilitate a major rework of the ways of working that we have been used to.

From our own experiences running an AI adoption programme within Implement Consulting Group⁴, as well as helping many clients tackle these challenges head-on, we believe adoption cannot be placed solely in the hands of either change managers or IT/AI professionals; the new AI era – the generative one – is truly one of close collaboration between IT and the business – and adoption is often the first test of that relationship.

The promise of generative AI and the adoption challenge

While previous automation waves have primarily impacted physical tasks and products, generative AI has set new sails, and the course is pointed towards transforming knowledge and white-collar work. Through embedded capabilities, generative AI can support creating, analysing and modifying text, speech or graphical content – and at the same time inspire users to engage, ask and brainstorm. We are seeing the boundaries of what can be achieved with generative AI be moved on an almost monthly basis.

The promise – for both organisations and their employees – is that generative AI can fundamentally transform work and business. By automating some mundane tasks and providing real-time, on-demand assistance for others, employees can spend surplus time focusing on areas where their uniquely human skills are most needed. This transformation empowers employees to dedicate more time to the meaningful, creative and strategic aspects of their roles. Or perhaps they can spend the time reflecting on personal growth, pursuing their ambitions, or simply recharge. Ultimately, the goal is to create a work environment where technology enhances human potential, fostering a healthier, more balanced and fulfilling professional life.

But where do these grand promises come from? Through following and assessing multiple studies over recent years, we maintain an overview of studies measuring productivity and quality benefits of generative AI within individual working tasks⁵. The studies indicate that individual tasks can be performed 32% more efficiently and with 18% higher quality. These numbers span own projects we have conducted with ambitious clients, numbers from a bold test of management consultants at Boston Consulting Group, as well as among white-collar workers in task writing or customer service settings.

However, while the numbers and potential are striking, we do not advocate adopting generative AI technologies for the sake of the technology alone. Despite undeniable potential, there are critical dilemmas and deeper questions embedded in these transformations. The challenge is not whether AI can make work faster or more efficient, but whether we are using it to fundamentally improve how we work. For instance, is AI helping us rethink outdated practices like unproductive meetings, or are we merely layering technology on top of inefficient systems?

No matter the stance on the big questions, future ways of working will not change overnight, and the de facto adoption rates of generative AI technologies show signs of caution all the same; the potential users of our generative AI tools in our organisations are not as blown away by new, fancy AI tools or promising productivity numbers as we and management are. Explaining this, we believe two conflicts are at play, and we label these under one term, “GPT hesitancy”.

We define “GPT hesitancy” as:

Personal conflicts: an internal struggle between reaping the promised productivity and quality gains to do more and/or better work, and the fear of seeming less skilled to one’s peers because using generative AI tools may be construed as “cheating”.
Inter- and intraorganisational conflicts: new tools, especially powerful ones, tend to bring about new ways of working that might sometimes challenge what the collective “we” are used to. This is “change resistance 101”, but the difference in the generative AI era is that other organisations – our competitors, peers and partners – can inadvertently influence our ways of working, too.

Our approach to generative AI adoption: Business and technology in duality

At the core of our approach to successful generative AI adoption is a people-centric mindset. Technology alone cannot, and should not, be expected to drive adoption – let alone transformation. The challenge lies in engaging humans: persuading them to change their ways of working, adopt new processes, build new habits⁶ and ultimately help realign the organisational culture around how to utilise generative AI tools.

While conceptual models are powerful in their parsimony, offering a simple framework to understand complex problems, they often fail to capture the full spectrum of assumptions, prerequisites and real-world variations necessary for successful implementation. Nevertheless, we believe it is important to present our approach, with your implicit understanding that it is a starting point – easy to grasp but harder to execute.

This duality – between business and technology, between simplicity in design and complexity in execution – remains at the heart of our generative AI adoption philosophy. We are dealing with humans, after all.

The six dimensions of successful generative AI adoption

We believe that successful adoption of generative AI tools relies on six key dimensions that are covered in detail in the following sections. As stated, the framework does not present a one-size-fits-all solution, and it comes down to individual organisations which dimensions need to be activated – each organisation’s journey is unique, and it's essential to tailor the approach to your specific context, goals, and readiness.

While the dimensions cover the approach, it’s equally important to acknowledge that change management underpins every aspect of AI adoption. The shift towards generative AI is a change journey, and ensuring smooth transitions, fostering acceptance, and managing resistance are all key to long-term success. Change management provides the foundation for these six dimensions, ensuring that they function cohesively to support the broader transformation.

Performance management

Following the mantra that “what gets measured gets done”, we have good experiences working with defining Key Performance Indicators (KPIs) or Objectives and Key Results (OKRs) in the scope of generative AI adoption programmes. KPIs and associated measurements and metrics provide a way to track and evaluate how effectively the adoption is progressing, ensuring that efforts are data-driven rather than based solely on intuition. It is imperative, however, that we measure the right things; for example, simply tracking the number of prompts used in an organisation or across teams might reflect the behaviour of a few, engaged users rather than providing a full picture of adoption across the organisation.

To get a balanced view, your measurements should capture both depth – i.e. how intensively certain users engage – as well as breadth – how widely the tools are being used across teams. Setting up such measurements and metrics should not be overly burdensome; instead, they should function as short-term indicators of your adoption strategy’s effectiveness. For example, you might track spikes in usage after specific communication efforts to gauge how well your engagement strategies are working. Over time, these metrics can serve as a foundation for assessing the true value and benefits that generative AI brings to the organisation, a task that is notoriously difficult to quantify, yet critical for long-term success. Ultimately, performance management efforts help assess whether the benefits of generative AI adoption are being realised and whether critical improvement areas are being continuously identified.

Recommendations

Set up clear, achievable indicators that reflect the progress of your AI adoption journey and consider both breadth and depth in your measurements. Do not rely on metrics that only capture a small segment of users; ensure your data reflects adoption across the whole organisation.
Keep your indicators simple and easy to interpret. Overcomplicated metrics can lead to confusion and make it harder to track progress effectively.
Regularly review and refine the chosen indicators. As both the tools and the adoption process evolve, your measurements should adapt to reflect emerging challenges and new opportunities⁷.

Leadership

It is critical that the leaders in organisations lead by example, showcasing how they, on a personal basis, use generative AI tools in their own daily tasks. When we see leaders actively using generative AI tools and openly sharing their experiences, we see that they help foster curiosity, trust and engagement. Such visible role modelling not only drives enthusiasm but also helps overcome scepticism, signalling to the organisation that AI is a strategic priority. You want leaders not just telling their teams to engage, but leading the way in doing so.

Beyond using generative AI tools, leaders must provide clear guidance on how their use fits with the organisation's values and strategic goals. It is crucial for leaders to define where generative AI adds value and where it may not be appropriate, ensuring alignment with business objectives. Leaders also play a key role in addressing ethical considerations surrounding AI use. Ethical leadership in this regard is vital to ensure the responsible use of AI across the organisation, while staying ahead of evolving legal and compliance requirements.

Recommendations

Encourage leaders to visibly adopt generative AI in their daily work and regularly share their use cases with the broader organisation. This will inspire others and promote widespread adoption.
Provide leadership-specific training, focusing on building AI competencies, understanding compliance guidelines, and staying up to date with relevant AI legislation.
Leaders play a key role in putting the topic of (generative) AI on the agenda, while ensuring it does not overshadow core business priorities.

Communication

Effective communication not only builds excitement and interest but also helps guide employees through the transition. Clear, consistent messaging can ensure that the entire organisation understands the value brought by generative AI, as well as how to make best use of it. From initial announcements to ongoing usage, communication needs to bridge the gap between training and everyday application, reinforcing key concepts and keeping adoption momentum alive. Consistent with the results of a survey by Gartner in 2023⁸ which found that 9% of organisations have an AI vision statement in place, an Implement Consulting Group (2023) survey found that only 6% of organisations had a generative AI strategy or team in place.

Well-crafted communication strategies or narratives also help address familiar challenges, such as hesitation and resistance to change. By consistently explaining the “why” behind AI adoption and offering ongoing support, communication efforts can help inspire curiosity and reduce fears employees might have about using new tools. Additionally, providing easy access to resources, such as tutorials, FAQs and on-demand resources ensures that employees feel supported as they begin to integrate AI into their workflows. This proactive approach can significantly enhance both short-term engagement and long-term success.

Recommendations

Maintain engagement through regular, digestible updates or learning tips to reinforce training and encourage continuous use. In Implement Consulting Group, we work with “Tuesday Treats” – the first Tuesday of the month, an e-mail conveys a new line of thinking or application of one of our generative AI tools.
It is beneficial to create a unified core narrative that explains why the organisation is adopting AI, ensuring all employees and stakeholders understand the value and purpose behind the initiative – and its link to the corporate strategy or vision.
Consider establishing a centralised resource hub (e.g. SharePoint, learning platforms etc.) where employees can access tutorials, guides and other supporting materials to facilitate ongoing learning and troubleshooting on demand.

Training

For AI adoption to truly succeed, thoughtful training and capability-building programmes must be embedded in dedicated learning journeys. It is essential to align expectations and equip teams with the skills and confidence to use AI tools effectively, but training alone won't do it– research suggests that 70% of professional competence is developed through on-the-job experiences, with another 20% coming from interactions with peers⁹. This highlights the importance of not just formal training, but also learning through experience and collaboration. But we are faced with a dilemma, because training is critical for sustained AI adoption, but at the same time training is in no way enough on its own – it must be combined with communication, supporting materials and strong leadership to help nudge and in turn consistently help change the habitual nervous systems in organisations.

A well-designed learning journey should challenge participants while fostering trust and relationships among colleagues. By sharing experiences and offering mutual support, employees gain valuable insights and motivation, which are crucial for long-term growth. While experiential learning and peer interaction form the bulk of skill development, structured courses – comprising about 10% of professional competence development – provide the theoretical foundation needed to enhance practical abilities. We recommend a balanced blend of face-to-face and virtual training to maintain access to learning materials without disrupting day-to-day workflows. Keeping training efforts relevant, updated and offered on a continuous basis ensures that it remains impactful.

Recommendations

Assess the specific capabilities required across various business functions and/or functional areas. This can guide which applications to focus on while guiding whether certain training modules should be mandatory or optional.
Create learning journeys with variation, e.g. workshops, inspirational keynotes, self-study materials and hands-on practice sessions. Variety is essential to keep the training engaging and effective.
If you struggle with finding time in the organisation to conduct training, consider incorporating microlearning – short, focused learning sessions that can be completed or watched in minutes.

Human-centred UX & functionality

When adopting generative AI tools – whether you are building, buying or already have one in place – the focus must be on creating solutions that are dependable, efficient and aligned with the needs of end users in your organisation. Tools that are overly complex or unintuitive risk alienating employees, preventing the organisation from realising the full potential.

During adoption programmes, the responsible group must emphasise and push for end user-friendly design, ensuring that tools fit smoothly into existing workflows and expectations. While the “adoption team” should not need to get involved in the technical development directly, they play a critical role in connecting end user needs with developers or representing them in procurement decisions. By incorporating multidisciplinary, human-centric UX perspectives, adoption is also about ensuring that tools are intuitive, accessible and effective for the intended audience. We jokingly say that the best thing our clients can do for their adoption efforts is to push out generative AI tools with organisational IT policies to the desktops and taskbars of all end users, and that this gets more immediate, short-term effects than any training ever would.

In the same vein, the functionality of generative AI tools plays a key role in user adoption, and an adoption team can and should engage in pushing for such considerations in development roadmaps or procurement decisions. If some >= 70% of users ask for unified capabilities such as useful image generation or organisational knowledge anchoring, this needs to be prioritised¹⁰. Of course, as changes or improvements are added, on-demand content and training sessions may need an update.

Recommendations

Aim to address 80% of your organisation’s core needs instead of getting bogged down with unnecessary, advanced features.
Put real users at the centre of every design phase. Consider their psychological, emotional and environmental contexts to ensure a better fit.
Partner with UX designers to create intuitive interfaces. What seems clear to AI experts may not be obvious to others.

Culture and ways of working

Organisational culture and ways of working – the centrepiece of the adoption model – function more as a product of the other five adoption dimensions. Culture is what happens “between the lines”, i.e., how leaders manage and inspire their teams, how the change is communicated, and more. When the other aspects are integrated effectively, a supportive and adaptive culture emerges, where generative AI tools are seen as a valuable helper rather than a threat, in turn shaping the collective ways of working. Culture, however, can also be influenced by specific organisational goals, targets, and how success is measured. For instance, incorporating AI adoption into employees' individual development plans can foster a culture of learning and innovation, but if not managed well, it could also lead to pressure or unintended consequences.

Here is how we work with culture and Ways of Working in a generative AI context at Implement Consulting Group:

We promote AI as a tool to enhance work, not as a shortcut
At Implement, we emphasise that our generative AI tools are designed to boost productivity and creativity, not replace critical thinking, human effort or strong subject matter expertise. We focus on establishing clear boundaries to prevent misuse, such as discouraging the sharing of sensitive information or overreliance on AI for tasks requiring human judgement and insight.
We foster an open environment for dialogue about AI
Creating an open, transparent culture is key. We actively encourage discussions around AI usage, allowing employees to voice doubts, share ideas and provide feedback. This approach ensures that AI adoption aligns with both our ethical standards and organisational values.
We leverage early adopters to drive cultural shifts
We identify and empower employees who are naturally inclined to embrace AI and position them as champions, e.g. through our “AI Champions” initiative. Sometimes these exist in the form of early adopters – for whom generative AI comes naturally. Their positive influence and expertise help create momentum for AI adoption across the organisation, inspiring others to follow that same line.

Other observations and key recommendations

Finally, before finishing off, we wish to convey a few other observations and key recommendations based on our experiences in projects with bold and ambitious clients who have embarked on the generative AI adoption journey with us by their side.

1. An AI strategy is essential

A clearly defined AI strategy serves as a guiding star for all AI-related efforts within the organisation. We’ve observed that many organisations begin exploring tools or use cases without fully understanding how they align with the business goals of the organisation. Establishing an AI strategy early, or pushing to get one in place, can help ensure a clear sense of direction and that both adoption initiatives and technical implementations are aligned with the organisation’s broader objectives. Without this foundational element, AI adoption can become fragmented, leading to missed opportunities or misaligned investments.

2. Governance is key

When it comes to governance, there are a few ways of organising a team in a decentralised or centralised setup. The most important thing, and our strongest recommendation, is to appoint a dedicated AI adoption lead – to ensure clear alignment between IT and the business, ensuring that the adoption efforts don’t become either/or, but rather both/and.

An adoption lead must drive adoption efforts, resolve challenges, align teams around common goals, such as coordinating with AI/IT teams on UX/Functionality on behalf of end users. Effective governance also allows for better resource allocation and more coordinated and consistent efforts across the organisation.

3. Awareness of the regulatory landscape is not to be underestimated

As AI technologies continue to evolve, so do the regulations that govern their use. It is essential that everyone in an organisation understands these regulations and their impact on daily operations. An adoption programme plays a pivotal role in ensuring that employees not only embrace AI, but also that they use it responsibly and in compliance with external legislation – often guided by internal rules, guidelines and policies. This is not only important for individual usage, but also crucial when making key decisions on which tools to implement or which use cases to pursue.

Regular training in and updates on AI-related legislation and compliance will help the organisation avoid legal risks and uphold ethical standards. As outlined in article 4 of the EU AI Act¹¹, “Providers and deployers of AI systems shall take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf (…)”.

Starting your journey

Many organisations have already embarked on the journey of adopting generative AI, and it is clear that employees are eager to leverage these tools to enhance their work. As AI becomes more integrated into and even changes ways of working, employees will increasingly look to their organisations for guidance and support in using these technologies¹² .

It is our belief that following the human-first approach, focusing on the five key adoption dimensions, you will not only meet this growing demand but also create an environment in which generative AI can truly thrive.

As stated, the benefits of AI adoption are vast, but success depends on how thoughtfully you manage to steer the process and guide your employees; they need to feel empowered and supported, and that is where your focus on a balanced, strategic approach makes all the difference.

Sources

^{1) Implement Consulting Group (2024) have estimated +8% GDP growth over the next ten years in the “D9+” group of countries; 12 digitally advanced countries within the EU: https://implementconsultinggroup.com/article/the-economic-opportunity-of-generative-ai-in-d9.}

^{2) McKinsey & Co. (2023), in an analysis of the Danish market and potential, estimate an expected GDP growth in Denmark of approx. 10%: https://www.mckinsey.com/dk/our-insights/det-okonomiske-potentiale-af-genai-i-danmark}

^{3) McKinsey Global Surveys, (2023, 2024): https://www.mckinsey.com/featured-insights/mckinsey-global-surveys and McKinsey & Co. (2024): https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/charting-a-path-to-the-data-and-ai-driven-enterprise-of-2030}

^{5) Chang & Noy, MIT (2023), Peng & Kannan, Microsoft Research (2023), Nielsen, Nielsen Normann Group (2023), Tabachnyk & Nikolov, Google (2023), Goldmann Sachs Research (2023), Chui, et.al., McKinsey & Company (2023), Dell’Acqua et al., MIT, Harvard, Boston Consulting Group (2023), Brynjoflsson, Li & Raymond, National Bureau of Economic Research (2023), Korinek, National Bureau of Economic Research (2023), Congressional Budget Office (2023)}

^{6) At the end of the day, we are trying to break the habit of how the “collective” we are used to working in corporate settings – e-mails, meetings, meeting minutes. Habits are not easily changed, and it takes a sustained effort to ensure good generative AI experiences can be transformed into de facto habitual changes.}

^{7) As an example, “advanced voice mode” was recently released by OpenAI and made available for real-time API integration via e.g. Microsoft Azure. Consider what you get from measuring the number of voice-based prompts versus decibel measurements in your offices. It is difficult, but we advise thinking long and hard about what you want to capture and how to best do that.}

^{8) HBR (2023): https://hbr.org/2023/12/5-forces-that-will-drive-the-adoption-of-genai}

^{9) Lombardo & Eichinger (1996)}

^{10) Read more about knowledge anchoring using RAG (Retrieval Augmented Generation) systems here: https://implementconsultinggroup.com/article/building-high-quality-rag-systems}

^{11) https://artificialintelligenceact.eu/article/4/}

^{12) More than 40% of employees expect changes in their roles or the need to upskill as a result of AI (HBR, 2023): https://hbr.org/2023/12/5-forces-that-will-drive-the-adoption-of-genai}

More on AI?

We are at the early stages of AI’s transformative potential on business, society and the future as a whole. But can we imagine where we are heading, and do we even dare? This is the challenge we want to overcome.

More on Implement's take on Artificial Intelligence