Most days of the week I’m busy working with a team of AI engineers Replacing mind numbing administrative work with automation tooling, (augmented by generative AI and machine learning) allowing professionals to get on with other more interesting and client-centric endeavours. One thing that stands out is that the AI projects I’ve been involved in all follow a pretty straightforward three phase pattern:

  • The Pilot phase
  • Scaling up phase
  • Consolidation and improvement

Disclaimer: I’m not talking about MlOps here, or any technical project challenges, this is very high-level basic stuff, that tech project managers seem to forget about when moving forward on innovation initiatives.

Each of these phases has its intricacies and challenges, and yes, I could dial down on the jargon and just call these three phases “making it work”, “getting more people to use it” and “making sure it doesn’t fall apart” but this is Medium, so I’m sticking with the jargon.

This is going to be one of several articles that tackles the challenges of scaling up innovative technology projects within large organisations. I’m not sure if it’s serendipity or dumb luck, but in my career I’ve been involved with people who wanted to push things a little bit further than the rest, entrepreneurs and pioneers who saw technology as a key to competitive advantage. Along the way I’ve learnt a few things and I’m going to endeavour to share them with you.

Let’s not start at the beginning

Congratulations, you’ve achieved the seemingly impossible. You have managed to get something to work that is completely and unmistakably, bleeding edge, innovative. And no, it’s not just about getting it to work, you got it to work In:

- A large organisation

- One filled to the brim with individuals and their personal agenda’s

- A labyrinth of technology solutions both old and new

- All held together by a gordian knot of rules and regulations

It might not change your life in the way that waking up at dawn to see the sun rise over a beach in Bali might, or make you laugh like that one special person in your life can, but you walked out of your office at the end of the day with a deep feeling of personal satisfaction and achievement.

Unfortunately getting something to work is not the happy ending that you were hoping for; it’s a fleeting moment on a long journey that is going to ask, (read: demand) that you go back to all the obstacles you had the first time round and this time “do it right”.

Which means, no more:

  • “It’s just a pilot phase, let’s plug it in and see what happens”
  • “This tech is proof of concept, we don’t need support”
  • “Skip the testing, it’s Friday let’s go straight to production”. (shudder)

The good news is, if you get this part right you go from: Feeling good about yourself and getting a pat on the back on a Friday afternoon. To: Actually making a difference in how your organisation uses technology to gain a competitive advantage.

Transitioning from just working to growing:

This may sound a little backwards after coming out of a pilot phase, but I’m of the opinion that the key indicator of success when scaling is: The level of integration within the framework of existing organisational technology governance and architecture.

Note: Integration is not detrimental to innovation, the two can mutually co-exist, you just have to be smart about it. If we deconstruct the word “integration” in a tech solution it falls into the following categories:

- Architecture

- Governance

- Information Security + Compliance

- Data Privacy

- Service and Maintenance

Successfully integrating the technology also means realistically calculating the resourcing around each one of these categories. Which, again, doesn’t mean you dilute the ambition of your solution, it just means that this phase is all about going through each step of integration appropriately and by the numbers.

To scale successfully you are going to need senior IT leadership time, you are going to need a solutions architect, the team leads at customer service and tech support are going to have to be involved, a DPA… the list just goes on and. Each one of these professionals wants a realistic appraisal of time, effort and complexity during the integration phase, not the high level promise of the technology or the business case.

So in a nutshell: You know that you are ready to scale when the initial effort and resources required to integrate weighs up against the current pipeline of positively impacted business processes.

The idea is that (in the reality of a large organisation) Once you’ve integrated correctly you can start working at scale.

It’s always about the people:

Without a doubt the best people to keep in contact with during scaling are the ones that pick up the phones when something goes wrong. Customer and tech support know exactly how good your new tech really is, both inside and outside the organisation.

Tech support teams are full of people that have, at one moment or another in their careers, worked (multiple) nights trying to recover data, spent weeks running and re-running performance tests praying for the right outcome, sweated bullets making sure that that online platform you promised to everyone is up and running by the morning. These people live and breathe the quality of your solution and most of them have a sixth sense for instability.

Customer service, are the ones that combine high level people skills with a generalist knowledge of your organisation, and they live within a context of dealing with the unexpected. They have answered calls from angry customers hour after hour, and stayed civil (no small task). If your tech hits the outside world and its not up to scratch, they are the ones that take the hit. The real world of your solution is right here.

Obviously I’m just scratching the surface here of the effort required to scale, in the next article I’ll zoom in on information security, the practical applications of certifications and how you can lock something down and let it grow at the same time.

Thanks for reading,

Kai