A Holistic Delivery Approach for AI: Part 1 - Introduction

Welcome

Welcome to the first in a five-part series where I will be outlining what I call the “Holistic Delivery Approach for AI”.

Across this series I will set out the rationale for why I believe that the industry needs a new approach for delivering impactful AI solutions and dive into the key capabilities required across the classic people, process, and technology framework.

This first entry will provide the context for this new approach and outline the four detailed areas we will investigate over the coming weeks.

So then, why do we need new approach for delivering AI? The answer is quite simple actually. AI has a ROI problem that needs fixing.

AI has an ROI Problem

The modern concept AI has been around since 1956 when it was first conceived at a conference of academics at Dartmouth College, but it wasn’t until the arrival of “big data” and the inexorable growth in computational power enshrined in Moore’s Law that we saw AI transition into a technology with practical business applications. If asked to put a date on when AI came of age, we could point to May 2017 when Google DeepMind’s AlphaGo AI defeated Chinese champion Ke Jie as a moment of pop culture recognition.

Putting aside any actual start point, AI has undeniably been a hot topic for business leaders with industry investing in AI technology in search of promised returns. However, after years of investment, thousands of proof of concept projects, and billions of dollars spent (US$119bn in 2022) we are struggling to find broad and compelling examples of where AI has delivered truly transformational change outside the very top enterprises in each industry. In fact, we are getting to the point where many leaders are questioning their investments in AI with. Recent research from McKinsey says that only 22% of organisations report quantified returns from and Investment with further research from Capgemini stating that up to 85% of AI project fail without bringing expected value.

The question then for us as industry professionals is to understand why AI is struggling to deliver a return on investment. Is it fundamentally not an applicable technology for most businesses or is it more a question of needing to adapt our delivery models and methodologies to fit with AI? Fortunately for us the answer is very much the latter. I believe that this ROI problem is primarily due to an inability to effectively deliver complete AI solutions that move from a POC to a fully realised business capability.

As the AI industry has matured over the last 5 we have observed and can now understand why businesses have failed to deliver on the promise of AI. We are now in a position to help clients avoid these common pitfalls and make AI a success. Based on years of frontline experience delivering AI transformation, and with the counsel of experienced industry colleagues, we have developed a delivery methodology specifically designed to ensure the effective end-to-end delivery of AI solutions. We call this the Holistic Approach to AI.

Essentially we believe that clients need to treat AI as a whole of organisation capability. They need to build all the necessary people, process and technology elements that power complete AI solutions. This approach is different to how we see many market participants addressing AI where they look to be creating new silos of expertise in a specific area such as MLOps or rapid prototyping or analytics platforms whilst failing to address the wider context in which these specifics areas need to work.

Without a holistic approach clients are at risk of having piecemeal capabilities which when combined are not sufficient to power enterprise ready AI products and services.

Unpacking our Approach

Over this series I will unpack the Holistic Delivery Approach across four parts:

1.      The Problem

There is no way to escape the fact that AI is difficult and presents its own unique delivery challenges. In recent years businesses have spent increasingly large amounts on data and analytics projects but in most cases struggled to get returns that deliver on the undoubted promise of AI. To demonstrate just why businesses have struggled with AI we developed a concept we call the “AI Scaling Chasm”. In this part of the series we will look at the common pitfalls of AI delivery and, more importantly, how to avoid falling into them.

2.      The Process

In this part of the series we will look at our overarching approach and process for value driven AI. To generate real business value from AI you have to follow a set approach that takes you from value based ideation through POC and pilot to a fully integrated business solution. Borrowing from proven methodologies in computer science and R&D we have developed an end-to-end process for AI delivery.

3.      The People

We also need to understand how to organise your people for successful AI delivery. Just like Digital Transformation ushered in a new “way of working” so too will AI. We have designed a specific delivery methodology for AI projects that enhances and builds on Agile principles – we call this Podification.

4.      The Technology

To deliver AI there are core technology foundations that need to be in place to support complete AI solutions. The current technology landscape for AI is bewildering with multiple providers all claiming to have the best solution be it a complete platform from a hyperscaler such as Microsoft or Amazon or much more bespoke offerings or point solutions for say data ingestion or model training. Within this space we will identify the two key areas of technology stack that are most critical for successful AI.

By putting all these parts together you have the core of the Holistic Delivery Approach for AI.

Looking Ahead

This concludes the first part of the series. Whilst I haven’t given away any secrets (just yet) hopefully you can start to see why we do indeed need a new approach for delivering AI. Looking forward, over the coming weeks I will set out in more detail the key underlying difficulties of AI delivery and how the Holistic Delivery Approach helps us overcome these problems.

I hope you enjoyed this first part of the series and look forward to bringing you more. In the interim if you’d like to find out more about the Holistic Delivery Approach, please feel free to drop a comment below or reach out directly. I’m always happy to discuss anything AI related! A