Enabling AI in your business:

Guest Columnist

July 1, 2024

Leveraging AI technology in the daily operations of a business can drastically change the way it functions – offering a myriad of possibilities.

However, it isn’t a one-size-fits-all, cut-and-dry process.

Integrating AI into your business has two parts:

Enabling AI in your existing IT operations
Deploying it into your business teams

This column focuses on a practical, how-to approach to accomplish part one – enabling AI in your existing IT operations.

IT management, growth recap

In last year’s The Business News column discussing IT management and growth pyramid, we identified the blend of IT members focused on the stability and utilization of your technology and those focused on the agility to explore and expand your reach and capability.

This bi-modal IT organization comprises of mode 1, which is operationally focused on efficiency and stability, and mode 2, which focuses on growth and agility.

This structure can vary based on organization size and industry technology demands.

Generally, your technology liaison or solution architect – the bridge between mode 1 and mode 2 IT – is the ideal person to introduce AI enablement.

They understand your technology ecosystem, the dynamics between mode 1 and mode 2 and the people best able to enable AI as an extension of current technology.

Extension is the key word here.

AI will not revolutionize your organization at its core but will expand your capabilities using existing resources.

Think of AI as elastic cognition.

This might seem like a foreign phrase at a business level but is familiar to IT.

Just as cloud computing components are designed to be scalable and elastic, AI should afford elasticity in cognitive processes.

This could include dynamic data enrichment, dynamic data classification and delegating certain control gates in processes that pass a certain tolerance – introducing consensus via a Retrieval-Augmented Generation (RAG) system.

AI components are essential because AI models evolve frequently.

You don’t want to strongly couple solutions to specific models but rather maintain extensibility throughout your technical organization.

The engineering behind mode 1, mode 2

The technical engineering supporting the IT organization has key responsibilities.

On the operational side, it’s about providing change management controls that minimize disruption, mitigate risk with toll gates and rollback procedures and standardize how changes are packaged.

This supports compliance and audibility within the organization, essentially encompassing DevOps and a continuous delivery pipeline.

On the innovation side, engineering accelerates workflows and the repeatability of work among creative teams.

For data scientists and analysts, it means enabling rapid exposure to data sources for experimentation, exploration and refinement.

This type of engineering simplifies and accelerates repeatable work and prepares the way to productionalize new data sets and machine learning models to support business, customer and reporting applications.

Though custom AI models and data scientists can be significant, most businesses do not need this level of customization now but should plan to enable it in the future.

The proof of value for AI

Using the crawl, walk, run approach to learning new skills, in this case, AI skills, should be applied similarly within your business – starting with a proof of value (PoV).

Don’t build skills for their own sake – always pair them with business needs and value.

Crawl: Identify valuable business problems and possible solutions using “thought at scale.”

Determine the AI capabilities needed and assess your organization’s maturity in these areas.

Identify the main drivers of success and failure in these solutions.

Goal: Create awareness around bringing new value into the organization via AI.

Walk: Commission a technical engineering project that builds capability, validates value and mitigates risk against the main drivers in the AI solution.

Prove it’s possible – what portion of value can be realized and if enough risk can be mitigated to introduce the AI solution into the organization.

Goal: Validate that the right amount of value can be realized with the right effort to justify viable change.

Run: Involve mode 1 and mode 2 IT teams to develop a full solution that brings the right changes with operational controls satisfying change management needs.

Create a solution that can be continuously improved based on metrics and feedback.

Goal: Introduce successful change and iterate on continuous improvement.

Where to get outside help

With new technology, you don’t know what you don’t know, and figuring it out costs time and money.

Investing in your team’s knowledge and experience is crucial, but outside help can help expedite this process.

Crawl phase: Utilizing external help to define what is possible with “thought at scale” or potential solutions can be beneficial. Service providers with many use cases across industries can spark ideas on AI solutions and their implementation.

Walk phase: Validate viability and value quickly with external expertise to focus the approach and means to measure success for your IT team.

This keeps your team focused and accountable to goals, leaving them with a positive experience for the future.

Run phase: Setting up repeatable and sustainable AI processes can involve a steep learning curve – therefore utilizing third-party services can help – such as Roli.ai for AI model abstraction and interface reusability; and Softly Solutions to stay current and relevant with governance and sustainability.

Conclusion

AI technology is best thought of as a component extending your existing technology ecosystem. Follow a process to link business value to your AI capability development.

Know your team, plan to invest in them and use external experts to guide the path in the short term.

By doing so, you can effectively roll out AI in your business, ensuring you leverage this powerful technology for maximum benefit.

TBN