With a plethora of AI related tech deluging the market, as an IT service, you'll no doubt be inundated with requests for software promising great outcomes and capabilities for the organisation. Combine this with senior pressures to embed AI as part of your business strategy and it’s likely you’re already thinking about when and how to bring AI in to your business.
Artificial Intelligence provides intelligent outcomes and automation to data it knows about, applying context to the information it obtains and performing calculations on the probability of (for example chat gpt) what you're asking for and what you want as an answer.
The critical notion here is that it's using either a large data set available to it outside of your organisation, inside your organisation or a hybrid of the two. If you want AI to understand what you're asking and for and stand a good chance of getting a valuable answer you need to start with clean representative data.
So where is all this data in your organisation and how do I get access to it I hear you say? That for many organisations is the $64million question. Providing a baseline of your data, where it sits and how it is used is often the killer question. Securing this data in use of AI also means understanding things like data sensitivity, data accuracy, data veracity etc.
Holding a central information asset repository is often the starting point for IT teams. But is this enough, just to know what you have? We think the context of how it's used, where it's used and who buys it are critical answers that also need to be factored in to ensure that data, people and technology work in harmony to create value for the organisation when adopting AI.
Getting a strong return on investment from any new technology adoption means assessing the value, benefits and complexity to implement. Here are the top five areas we’d recommend exploring before making a decision on adopting an AI based technology.
Understand the use case - What will this new AI product be used for and what technology, service or process does this bring additional effectiveness or efficiency to? Who will use it and how valuable is it to those business services, employees or customers. Getting this insight through mapping this use case to existing understanding of business services, capability or customers in a particular part of the organisation is critical for understanding ROI and how / where it fixes a particular problem.
Understand what is used currently - What technology in the organisation is already offering a similar level of benefit? Is there a product that could be extended through use of AI to make it more effective, easier to use and aligns with existing regulatory and security standards? Exploring what you have, where it’s used and why it’s needed could help you avoid cost and make the adoption experience easier for end users.
Understand technical dependencies of adoption - What data will this new technology use that makes it relevant to your organisation and helps with the outcome you're trying to achieve? What end user equipment will need to be provided for end users to consume the outputs from the AI solution? What internal applications and data sets will it need to be integrated with to ensure you get the most benefit from it? Exploring technical dependencies helps you understand the complexity to implement, the cost to implement and what the impact on support might be in the future.
Understand the risk to adoption - What is the sensitivity of the data that the AI solution might need to integrate with and how does the AI solution protect this if it is accessing this? Is it learning from this data, that ultimately then means it appears in another organisation’s results from queries? How is it protecting customer data, sensitive data or intellectual property your organisation owns? Is the solution going to be relying on out of date or stale data? It’s critical to understand the different data required, their sensitivities to the organisation and how this is protected by the solution being put in place and align to the outcome / outputs of the solution.
Understand the IT service and business dependencies - Do you have the right skills and capacity to support end users with this technology? Does the organisation have the right skills to use and adopt this technology? Is there any risk to the organisation from a culture or reputational perspective from hidden biases or hidden ways in which the AI solution is working / producing results? It’s key to work with the vendor, your IT team and the business to understand these elements, ensuring you have the skills, capacity and it fits both within the culture and way of working within the organisation without bias that might harm the reputation from decisions made through use of the AI tooling.
A baseline of understanding between technology and business really helps in all of these examples. Having the relationship to hand between the likes of data, reporting and applications provides immediate insight into the context of why data may exist. If you look at secondary relationships between the likes of applications and reports consumed by business services or organisational units, you can start to map how data is used and by who.
All of this provides valuable insight on which to investigate further or spot opportunities for change there and then. This baseline can easily be surfaced through the likes of CoPerceptuo. It provides the collaboration tools and central data repository to harness the right information needed to enable this insight.
If you'd like to understand more and see what other possible context and sights you could add to help you in building your AI strategy get in touch here.