Building a strategy for artificial intelligence

As we approach ChatGPT’s first anniversary on 30th November, Mark Wagman at MediaLink asks, Who owns your organisation’s AI strategy? (Clue: It starts with you, but bring some friends)

 

Artificial Intelligence (AI) and machine learning technologies have been commonplace across a number of industries and enterprise functions for years. But up until recently, it played more of a behind-the-scenes role for more large-scale organisations. 

 

This next era of Generative AI has changed that, bringing front-of-house use cases to life before the eyes of a boardroom. What was once just a bunch of 1s and 0s is now taking (and passing) tests to become lawyers and doctors. As with similar disruptive technology, with such opportunity comes risk, raising the question, across many leadership teams, “Are we as an organisation even ready?”

 

As some boardrooms swirl, and others face mutiny and disruption, wasting cycles chasing the flashy AI use cases that make the trades, sophisticated leadership teams are taking the time to plot a path forward – one anchored in creating opportunities to learn and evolve quickly.

 

Key questions leaders tackle at the outset include: what does your business want out of this next technological revolution? What does your organisation need to reach its next level of growth? How can we actively manage AI integrations to avoid our version of Space Odyssey? What are the ethical and societal responsibilities that will fall on the shoulders of businesses with these new superpowers? 

 

The most successful leadership teams must first be honest with themselves and their teams, admitting that the path forward is unclear unless they take the time to set the course. In the case of solving AI’s role in the enterprise, that starts with defining a core set of use cases to explore and taking that first step on a winding path forward.

 

A cross-functional team that understands the business

The transformational opportunities created by AI can only be realised when you expand your thinking to include input from colleagues and leaders across the organisation (and across all levels). Yes, the AI strategy starts in the boardroom, but it is very much built from the ground up by leaders from across functions, business units, and roles.  

 

Teams should assess where friction exists in the business today, where it is spending disproportionately more man-hours and why, who manages the largest datasets, where the primary customer touch point is, and how to make that a better experience. 

 

Increase the likelihood of success – the likelihood of adoption – by bringing together voices from across your business functions. An AI strategy must be pragmatic and rooted in key use cases that make the company faster, more efficient, and more effective –– those fundamental building blocks start and end with your teams. 

 

Define a collection of destinations: create pilots and learn

The next step leaders should take is to set a path –– or series of paths –– forward, and then, and then fail quickly.

 

Define a handful of use cases or a collection of hypotheses, and get started. Ask yourself: where can we accelerate efforts that are already working? What manual processes are ripe with errors that could benefit from AI assistance?

 

In my own media and advertising industry, marketers may look across productivity, consumer experience, and growth initiatives for opportunities to inject AI into processes ripe for acceleration and efficiency. 

 

Can we create hyper-personalised eCommerce experiences for our customers, leading to higher conversion rates and, ultimately, sales? What can we do to decrease paid customer acquisition costs and find a higher “lifetime-value” consumer?

 

Those at the forefront of the revolution, such as Alibaba in Asia, are using AI to not only predict what customers might want to buy but also automatically generate product descriptions for their site using natural language processing. 

 

The lesson? Define the pilots –– get specific on the inputs and desired outcomes. And remember that you can’t solve it all –– so avoid “paralysis by analysis” –– and look for patterns and trends. 

 

Make friends; create partnerships

Don’t go at this alone. Don’t pretend that overnight; your executive team all received data science master’s degrees from Stanford. Go find the experts – seek out the partners and people who know what they are doing. 

 

Start by looking within your vertical or use case and explore partners using AI to optimise thousands of creative permutations to find the right fit. Are you aiming to streamline accounts payable efforts to get money into the hands of your vendors more quickly? Consider platforms using AI to reduce reconciliation efforts and cut down time-to-payment by days.

 

Iconic brand Pepsico identified from an early point the transformational impact AI could have on its business. By partnering with key business lines’ leaders, Chief Strategy and Transformation Officer Athina Kanioura is able to apply AI to all corners of the business.

 

Whether for research and new flavours development or to drive workplace safety in large-scale factory facilities, getting AI “right” at a global brand like Pepsico requires partners, coordination, and a lot of learning.

 

Think human first

AI will undoubtedly have an impact on workers, whether it’s a case of cutting labour hours or retraining workers and providing opportunities for the building of new skill sets.

 

A study from OpenAI, the company behind ChatGPT, found that around a fifth of workers in the US could see 50% of their tasks affected, for better or worse. Other recent studies have shown that generative AI is much more likely to augment work rather than completely replace jobs.

 

However, a level of handholding is still a necessity when using these tools and eradicates the argument that AI can operate independently without the need for the human touch - a lesson most recently learnt by two New York lawyers who were fined after submitting a legal brief with fake case citations generated by ChatGPT.

 

Training and retraining to bring everyone “along” is crucial. Ownership around a strategy needs to be granted widely, not in silos, and collaboration with outside partners will be paramount. 

 

AI is a technology that will fundamentally impact not only every function of an organisation but also the larger ecosystem in which it lives. So align with internal stakeholders and ensure the organisation is ready for AI and the implications of using the technology across mission-critical processes. 

 

And be ready to be there, step-for-step. Human assistance is still a critical part of this puzzle – the machines don’t get to have all the fun just yet…

 

---

 

Mark Wagman is Managing Director at MediaLink

 

Main image courtesy of iStockPhoto.com