April 22, 2024

AI: Why Now?  

How to Get Started With Minimum Risk

With the latest advances, it is now clear that AI is very much in the mainstream. Many of us are using AI embedded in our desktop and mobile apps for everyday tasks, and we can see that leveraging the same technology in our business applications can provide game-changing benefits.

However, it can be difficult to identify where to start, which use cases to work on first, how to build your own AI capabilities, and how to do that while maximising ROI and minimising risk. It can be all too easy to postpone making investments and adopt a 'wait and see' strategy.

The Risks of Waiting

The risks of waiting could be substantial. As soon as a competitor develops AI-powered, personalised front-ends and intelligent processes, outperforming your organisation on service, speed and cost, it may take too long to respond if you haven't already started your AI journey.

This is because successfully embedding AI into business processes takes time: it requires foundations to be built right across the organisation. To successfully implement AI will require an AI vision for the whole organisation grounded in the business strategy, updated technical architectures, new applications, education on all aspects of AI, and transformation to agile practices in business areas as well as in IT.

Whilst putting these foundations in place takes time, not getting started now presents a risk in itself: once a competitor is edging you out of your markets, it may be too late to catch up.

A Clear Path Forward

Fortunately, there is a consensus on how to get started. There is a clear stepwise approach that ensures the foundations are put in place for successful AI implementation and that AI use cases are chosen carefully to maximise ROI and resonance with the business strategy, thereby minimising risk and cost.

In the March Finativ Technology and Innovation Forum, I shared my experiences and insights on the essential foundations for successful AI implementation, including tips on how to get started, use case selection and lessons learned.

Key Success Factors

At a high level, the key success factors for AI are strategic alignment, technical readiness and organisational readiness.

1. Strategic Alignment

For strategic alignment, it is important to have a digital vision and an AI vision in place from which an AI roadmap can be developed. I outlined how I have gone about this successfully in the past by what I call a 'top to bottom to top' approach. The resulting vision should have elements matching the business strategy but also practical details that every function and member of the organisation should recognise and happily buy into. I'd check this by playing it back to the leadership and select groups, for example, a group of salespeople at one of their regular meetings.

Once an AI vision is established, it is best practice to have an owner in the business responsible for the wider transformation, and also a counterpart in IT. It is also advisable to have a specific AI leader to ensure regulatory compliance and cover the use of AI outside of the core business applications, for example, in desktop software. Finally, there should be a management system and steering committees in place, and also frequent communication from the business leadership on progress, providing clarity and demonstrating engagement and commitment.

2. Technical Readiness

Once a digital and AI vision is in place, a high-level roadmap can be established with business working together with IT. The digital and AI vision will drive the technical architecture and roadmap. The technical roadmap will generate pure IT items, such as an incremental migration to a cloud or hybrid cloud environment, and it would define a broad timeline for setting up a new or updated technical architecture. Both the business and IT elements should be reflected in the overall digital and AI roadmap. I emphasised that having strong business-savvy architects is absolutely critical.

Agile and DevOps

As well as creating and starting to execute to a roadmap, it is important to move IT teams to agile ways of working, and also implement DevOps and SecDevOps. Once AI is up and running, it is important to implement MLOps. Teams working on IT applications should be formed as agile mixed-discipline squads. This should also include legacy applications, in order to avoid legacy areas becoming bottlenecks. All the time, teams should share learning and code patterns and, at all costs, avoid silos.

Leveraging Existing Resources

If your financing business sits within a parent group, then look to leverage any AI centres of excellence and architecture if available. If the parent architecture isn't quite what you want, invest your time to influence it - resist the temptation to go it alone. Using existing AI expertise, platforms and resources is important to reduce cost and risk, and speed up implementation.

Building Internal Capability

It is important to build your own technical AI capability. Consciously procure and launch education on agile, design thinking and AI for all of IT, regardless of whether they are directly involved in an AI project. The aim is to have the whole organisation moving forward, bought in, knowledgeable and supportive.

A particular area at the outset for which there is no straightforward approach is determining the balance of internal vs external AI capability, in terms of people and technical environments. It does depend on the existing level of any parent support, existing tech partner capability, the existing IT environment, and the complexity, scale and posture of your organisation. I would say that to reduce cost and risk, it is necessary to avoid going to one extreme or the other at the outset, i.e., avoid doing everything internally, and avoid going for 100% outsourced.

Once established, AI will likely be difficult and expensive if 100% outsourced with no internal AI development environment. Therefore, a pragmatic approach would be to have some internal capability from the start but leverage whatever parent or external assistance is available as much as possible to start with and then migrate to have more internal capability over time.

Scalability from the Start

In choosing initial projects, make tech choices that can scale from the outset. Demos and experiments are fine, but it is important to flush out technical, skill or cost challenges as early as possible by using the scalable set-up from the start. Creating a technical environment upfront that can support multiple use cases is a way to make it easier, less expensive and lower risk to demo, experiment and provide proof of concepts. Future-proof to a certain extent by designing modules to allow for future swaps of technical capability.

Bear in mind that retrofitting a database, for example, will be expensive regardless of design, so make scalable decisions right from the start. Similarly, because a native cloud stack is so complex compared with a legacy equivalent, and requires constant maintenance to address compliance, dependencies and vulnerabilities, it is expensive to keep a cloud app in 'lights-on' mode. This means that if an application is not providing the value intended, the option of keeping it running with no development can be unattractive, and porting to another framework, database or tech stack may not be as straightforward as might be imagined. So, in summary, whilst it looks easy to stand up an application quickly, resist the temptation for any shortcuts – anything you create may be with you for a long time. That said, remember also that by selecting use cases which are on the AI roadmap the risk of benefits being smaller than envisaged is made as small as possible.

3. Organisational Readiness

It is crucially important to have visible leadership buy-in and engagement for the digital and AI vision. This can be demonstrated by providing regular communication on the vision at the outset and then frequently as the transformation progresses. It is important for leaders to attend steering committees, and that steering committees comprise subject matter experts, product owners and stakeholders as well as functional leaders, creating an open team ethic with one common purpose.

Product Ownership

A key consideration is to ensure strong product ownership. The product owner must have deep knowledge of the business and credibility with the wider business leadership, as well as be capable of leading the IT team and being able to address deep and complex technical challenges, and being able to handle mundane but important compliance and maintenance. Such resources are very highly valued and should rightly be the focus of retention and professional development initiatives.

One effective approach for widening the product owner resource pool is to appoint both a business product owner and an IT product owner of equal standing and working shoulder to shoulder. This can play to the strengths of both parties and can also make it easier for product owners to cover more than one application, thus potentially enhancing communication, interaction and efficiency across a programme.

Education and Alignment

Across the wider organisation, great benefits can be realised by providing education on agile, design thinking, digital transformation and AI for the whole business. If not the whole business to start with, then at minimum the business leadership team, subject matter experts, thought leaders in the business, and members of product or project teams.

If the IT team is agile, but the business is not, then it will be difficult to deliver and sustain AI projects. Similarly, if the business is agile, but the leadership team is not on the same page, then there can be disconnects which can erode trust, progress and delivery.

In particular, given the nature of AI, such projects can only thrive if there is trust, and a prerequisite for that is understanding.

How To Get Started

The most important act in getting started is to actually take the first step, having ensured that whatever you are doing will deliver the greatest benefit with the least possible risk and cost. The primary point is to make sure that the Strategy-Technology-Organisation success factors are satisfactorily addressed, especially the vision and roadmaps. If any one of those areas is weak, then it may compromise the delivery of AI projects or the programme as a whole.

The next point, following on, is to move from use case demos and experiments to items on the roadmap as quickly as possible, remembering to use the scalable technical set-up from the outset once you are beyond experiments. Select projects which are simple achievable AI use cases first, probably internal, and which have strong business SME support.

Practical Tips

Here are other tips based on what I have learned through creating and applying a digital and AI vision in asset financing:

  • Do leverage any internal and external AI expertise. Engage with and influence a parent group AI centre if it is not giving you what you want
  • Do try to use existing proven AI products at the outset rather than building your own from scratch. Don't just rely on a demo, but additionally, go and visit another organisation that is actually using that solution
  • Do check data out before kicking off a project: in all AI solutions, core data is likely to be important, certainly for the AI, but recognise that users require wider information for context to make AI usable, so check what data you have and where it sits
  • Do educate the target SMEs in AI, agile, design thinking and testing
  • Do support and protect the business SMEs, e.g., automate testing
  • Do empower and champion the business SMEs as far as possible, as they are the single biggest success factor

And finally . . .

Do ensure that the senior leadership demonstrates commitment and engagement by being seen to be involved, by communicating widely and regularly, and by empowering and championing the SMEs in the business.


Embarking on an AI journey can seem daunting, but by following a clear, stepwise approach that addresses strategy alignment, technical readiness and organisational readiness, organisations can lay the foundations for successful AI implementation. However, it can take time to put those foundations in place, so starting as soon as possible is advisable. Then by carefully selecting AI use cases that align with a digital vision driven by the business strategy, risks and costs can be minimised and the maximum ROI delivered.

If you're ready to start your AI journey or want to learn more about how to prepare your organisation for AI success, visit our AI Readiness page.

Simon Potts Finativ

Simon Potts

Simon Potts brings a wealth of expertise and experience to the field of AI and digital transformation.

With over 20 years of experience at IBM Financing, he has been the global lead on AI and digital transformation. Simon is recognised as an accomplished technology leader, holding global roles covering IT strategy and digital transformation.

Subscribe to INSIGHT

Subscribe to our newsletter to receive each issue directly in your inbox. Unsubscribe at any time. View our privacy policy here.
Subscription Form

Related Posts