A Practical Guide to AI Selection and Maximising Adoption
In our most recent blog we spoke about the importance of IA before AI (see blog here), but once this has all been setup for your needs the next critical decision is in AI selection. This has become a key strategic decision to avoid hype-driven investments, large costs and low or complete lack of ROI. The sheer variety of tools on offer all seam to provide clear advantages, like streamlining workflows, enabling automation, offering increased productivity, and even predictive analytics. But as teams adopt these at record speed, visibility, accountability, and cost management often fall behind.
It’s important for organisations to not get carried away in the momentum of AI and instead focus on business-driven initiatives. Here, we outline how to select and adopt AI tools for you, while making strategic budget commitments aligned to the organisation.
Road mapping AI selection instead of rushing
Where can AI add value?
It’s beneficial to take a step-by-step approach when considering AI selection for your business. Dig beyond the appeal of simplifying tasks, there is guarantee they will work for everyone. It’s important then to first scope out the potential need and clarify the specific business problem you’re trying to solve before moving forward with selecting a technology.
Workshops offer an effective entry point to start conversations regarding new initiatives. Bringing people together across business functions is a powerful way to immediately gain insight on the programme wide challenges that different teams are trying to solve. It also highlights whether AI would be beneficial business-wide or only for certain departments. These sessions can be designed to encourage brainstorming around potential use cases and desired outcomes, while ensuring alignment with the organisation’s overarching strategy and long-term goals. To get this right it’s important to allow every team to have a voice. Utilise techniques like round robins, dot voting, and breakout groups. The aim is to conclude if and where AI is most useful and to highlight and be aware of any constraints associated with this.
Following this discovery period, high-level plans for the procurement and process of implementation can be developed.
Cost Considerations with AI Selection
What should be considered when selecting AI software?
With so many tools out there, as well as personalised variations on these tools, it’s often overwhelming for procurement teams to identify which AI software is right for the business and delivers a strong ROI. Budget is a strong driving force in selection so here are some cost-related factors to consider when evaluating the options.
Costs associated with AI-powered tools comprise several core components and can be complex, including base licensing and subscription fees, token usage and computational costs, as well as ongoing maintenance costs to sustain performance and value. Work with your team or an external consultant to understand which factors are most important to your budget. This will begin to narrow your options to finding what suits you best.
Businesses should now have a rough estimate of how many licenses will be needed across the organisation based on prior internal workshops and discussions around AI selection. However, not all users may require premium subscriptions, advanced functionality, or add-ons. Mapping this out allows you to compare your options by calculating total costs across different capability levels to achieve the best ROI.
Let’s expand our understanding of token usage. Tokens are the basic units AI systems use to interpret language. They may represent full words, fragments of words, single characters, or punctuation. When a prompt is submitted, it’s divided into a series of tokens, and each one contributes to the overall cost. This is why scoping usage is essential prior to selection. If certain departments are only using it to perform simple tasks such as drafting basic emails or summarising short documents that offer little efficiency gain, then it may not be the best use of investment.
You should now be clearer on whether you require a personalised or out of the box AI solution. It’s important to be wary here of quick out of the box solutions. For example, many ERP solutions now offer AI capabilities, and it can be tempting to skip extended demos and purchase them as part of a bundled package. However, does the additional cost deliver meaningful value? Does the solution improve productivity and meet the intended use cases? Or is it simply a chatbot drawing on existing data that requires ongoing management while delivering limited practical benefit for your specific needs.
AI Adoption and Accountability
How to successfully adopt AI solutions?
A great starting point for AI adoption is creating a central, cross-functional team to oversee software requests and approvals. This team will be in charge of establishing strict procurement policies and processes for any new software including possible AI opportunities. This allows for immediate visibility on any new business-wide ventures.
Within these procurement approval processes, businesses can incorporate strategies such as pilot programmes prior to full deployment, helping to validate business value. An excellent example of this is through our work with Ridge (full case study here) to plan, coordinate and deliver a pilot of Microsoft Copilot to select business areas. The project delivered many valuable outcomes including the ability to validate value through feedback and tracking engagement. This approach informed an optimal deployment strategy and established a foundation for tailored communications, training, and onboarding for full-scale execution. This method of AI adoption is effective in keeping engagement and morale high, especially as new AI opportunities emerge from the successful implementation of others.
Returning to the idea of a centralised team to triage AI adoption initiatives, this is critical for enforcing strong governance, preventing shadow AI, and ensuring all efforts are managed through a portfolio-level view. Shadow AI refers to using and adopting AI tools within an organisation without formal approval, usually to solve immediate problems. It is important NOT to fall into this trap. This rapid adoption approach may speed things up in the short term but greatly limits visibility and introduces risks related to data security, compliance, and escalating costs. In essence, a business is back to square one, swept along by market momentum and fast-paced investment without a clear sense of direction. Clarity over needs, usage, budget and personalisation is critical.
AI Value Add
For CIOs, CTOs and IT directors all of this means identifying areas where AI could add value across the business, establishing strict procurement policies, and evaluating costs based on staggered licensing to maximise ROI. This ensures visibility, accountability and successful adoption of new AI solutions. Get AI right and enjoy the benefits in the long run. Chat to one of our team today to get your AI programme started on the right track.
