11 Jun Carving the Pathway to AI – Maturity & Business Readiness
By Dave Luttrell (AI Capability Lead | Principal Consultant)
Generative AI has caused an interesting trend in the landscape of AI and Machine Learning; it’s exploded into the hands of the users. Everybody is scrambling to fill the whitespace left by AI, hyperscalers are going all in, boards are demanding the latest and greatest. It is a bizzare and interesting innovation landscape.
This is underpinned by an interesting undercurrent and market perception, firstly, Gartner have found that “80% of executives think automation can be applied to any business decision.” A interesting statistic, particularly not considering whether it is applicable or appropriate to use AI or Automation for this purpose. To balance perception with a dose of reality, as at the end of last year it has been found “1 in 7 AI initiatives destined for success.”
Scaling AI is Hard
And indeed, it is hard to scale AI initiatives, and when adopting, or veering away from their maturity level, several organisations uncover innate complexities when rolling out and deploying AI based solutions.
- There is the Risk of Shadow AI, if you don’t provide sufficient platforms internally, employees will start finding the path of least resistance to make their lives easier. This provides a significant risk of Data Breaches, Leakage, Prompt Injection and Silos.
- Secondly is around that of Automation Complacency & Black Boxes which can envelop, particularly when it comes to risk sensitive decisions. AI fundamentally needs to be explainable, and we can’t just raise a defence to auditors or regulators saying, “The AI says so!”
- Why AI? – alluding to the 80%, a major fallacy I see is organisations trying to crowbar in use cases. We need to take a step back and think, what is AI good at… approximation, summarisation, generation, generalisation, and is this the best fit solution or recipe for the problem space we’re looking at. Sometimes the deterministic or rule-based approach, can be just as good, simpler and easier to implement.
Directing Your Focus Aligned to Maturity
An important place to assess your ability to navigate and institute AI is assessing where your organisation sits on the from an AI Maturity standpoint.
- If we look at those that are Hyped or Dabbling, they’re probably going through adhoc-pilots, quick wins, but are suffering infrastructure gaps, POC’s are piling up and probably have limited or no guard rails to support AI
- Those Enabling or Operating, have built solid foundations, have worn the scars of deploying models to production, or are bolstering deep seated capability to enable AI. They’ve also good oversight of the pitfalls and risks AI can enable.
- And looking at AI Native, I think this is a state that very few organisations will achieve. These are organisations that will enable AI to completely change how they achieve their value proposition
So, the question at this point is where do you place your chips when it comes to AI, do we go big or remain lean?
Going Big or Lean?
The first questions or checklist is to understand:
- Do we have a mature AI Governance in Place, which can support the non-deterministic nature of AI?
- Have you quantified the upside, and do you know how to measure success?
- Do you have the pipelines, ops and change management muscle to put models into production and keep them there.
If you can tick all three boxes, it’s usually worth placing a bigger bet — the scaffolding is there to turn ambition into value.
And the data backs this up:
- McKinsey found AI programmes with a P&L metric are 2x more likely to scale successfully.
- BCG have found through a study that only 17% of firms with weak governance managed to deliver value from AI vs 41% with strong governance.
- Firms with a mature state of AI/MLOps release AI initiatives 5x faster and cut failure rates in half.
How Do We Operationalise Key AI Support Structures & Initiatives?
A maturity model tells us where we are; this operating model tells us how we move forward — one disciplined gate at a time.
Domain Assessment
We start where the problems live — Finance, Sales, Risk, Operations. Each domain owns its backlog of business pain-points and opportunities.
Ideation
Before we jump to AI solutions and Models, we ask two questions: What actually moves the needle here? Which tasks are ripe for automation?
- Strategically — revenue growth, socially impactful, brand differentiating.
- Operationally — what’s annoying, low value, repeatable
Cross-functional squads translate those pain-points into AI concepts. Think of it as a design-thinking sprint:
- What Data do we have?
- What AI Approach might help?
- What outcome would prove success?
Evaluation
Hard-wire quantifiable outcomes and bring in the experts to have a rounded perspective on AI
- Eliminate cost $ per transaction, −15 % in year one
- Create benefit Net new revenue or social impact (e.g., fraud prevented)
- Enhance awareness NPS uplift, employee adoption, risk alerts surfaced
Pick the metric before building the model. That way every design decision traces back to a tangible outcome
Ensure initiatives check the box from an ethical, legal and policy perspective
Governance
- Ensuring there’s appropriate sponsorship and endorsement at an initiative level, while also aligning to enterprise policy and risk appetite.
- That discipline converts enthusiasm into sustainable value and keeps us out of the 85 % failure club.
Establishing AI Value
If we don’t manage cost, value and ROI deliberately, the default pattern is high sunk cost, low quantifiable value, and outcomes no one can quite trust.
Alongside an operating model we’ve articulated a playbook in terms of how they can measure and articulate value from their AI initiatives.
Most Important – Remember the People
“Models don’t deliver value; people using models deliver value.”
So, we validate usability, trust, and change-management at the same weight as precision and latency. A technically perfect model abandoned by users has an ROI of precisely zero.
LEARN MORE
To learn more about AI Governance & AI Use Case Prioritisation frameworks, please reach out for a conversation with our AI Advisory team. Start your AI Transformation today!