Outcomes of investing successfully

Outcomes of investing successfully#

When you’re writing the business case, it’ll help to set some waymarkers of what Leaders should expect to see after 12, 18 or 24 months (or beyond). This is really hard with applied AI (if that’s a surprise - go back to AI is changing everything), as the landscape is moving so fast, it can feel difficult to state exactly what benefits you will or won’t have targeted in 3 months, never mind two years.

However, asking for directional visibility is not an unreasonable question when you’re about to potentially spend millions of dollars! The best way to answer this question is to focus on what the conditions to do AI in the business will look like, and confidently predict the aggregated value delivered across a range of use cases, rather than focus specifically on an exact number generated from a use case long list that will be out of date in a year.

At the end of 12 months, what should you look around the business and see?

An AI-informed, educated and motivated team...

...on the road to becoming AI-native.

We’d expect to see:

1. Regular education events for leaders
2. A rich backlog of tested ideas, and self-served prototypes, of AI applications (developed by the business for the business)
3. Colleagues consistently challenging the art of the possible with AI
4. The business strategy has evolved towards an AI business strategy
5. A good understanding of the regulatory and ethical impacts of using AI within the business

Federated access to a good AI productivity tool to drive innovation

Everyone in the business gets the opportunity to become AI-native and build their own, meaningful, solutions with general-purpose AI.

We’d expect to see:

1. Deploy a task productivity assistant (like ChatGPT, Copilot or your own 'wrapper' app) ASAP
2. Set up non-functional logging on user activity, guardrails for topic moderation and appropriate data security and confidentiality nonfunctionals
3. Train people to use this technology; how to better prompt, construct workflows and experiment in their own time (see blog on adopting ChatGPT for enterprise)

A platform for taking AI to scale in your business

The last piece of the jigsaw is establishing a platform that will take the prototyping, or small-scale prompt-based solutions, to a meaningful production scale.

1. An initial, production-strength, AIOps capability able to facilitate: model management, observability, cost management, safety, performance and end-to-end testing
2. A significant percentage of the organisation’s knowledge is curated to be AI-compatible
3. Several applied AI reusable capabilities are prioritised based on those required to scale the most meaningful solutions: capabilities such as fact extraction, knowledge modeling, explainability lineage and retrieval to name a few