This week, in Boston, USA, IBM held their annual IBM Think conference. As ever, it promised to be a week of groundbreaking announcements, and it didn’t disappoint.
Just from the opening keynote from IBM CEO Arvind Krishna, entitled ‘Unlocking the Full Value of Enterprise AI’, there were some significant takeaways.
The focus of this keynote was how AI, automation and hybrid cloud are now critical drivers for enterprise growth, productivity and competitive advantage.
AI has moved on from the experimental stage to delivering real-world results; the hype is fading and people are focusing on the application of the technology. Now AI, particularly when used in combination with hybrid cloud, is leading to greater productivity, faster processes, reduced costs and improved security.
The theme of combining AI and hybrid cloud ran through the keynote like a golden thread. It forms the answers to the questions:
- How do I bring intelligence to all the operations I have?
- How do I bring automation to all enterprise workflows?
- How do I speed up decision making?
- How do I scale my business without adding a huge physical footprint?
Focus on enterprise data
There remains concern about how external data might influence business decisions. We have heard stories of hallucinations resulting in confusing, sometimes amusing, sometimes dangerous outcomes.
These hallucinations are inevitable if you rely solely on external data for business decisions. Krishna stated that 99% of enterprise data (that stored within an organisation) has been untouched by AI.
What this means is that there is an untapped resource of information that is specific to your organisation that will help your processes, your efficiencies and your decision making.
Of course, we’re not saying that you just upload this data to an open AI tool. This is a definite no-no.
So how do you unlock the value of this data?
Smaller is better
You need an AI approach that is tailored to your business. Rather than relying on LLMs, use special purpose-built models that are much smaller and created for a particular use case that will ingest the enterprise data and provide the answers you need.
There is data on the AI leaderboards that show that these smaller models are more accurate than larger models. After all, they are trained on a discrete source of data rather than being influenced by spurious data sources.
Furthermore, these small language models are faster, more cost effective, can run wherever you like – on premise, private cloud, public cloud or on a hybrid environment and are easier to integrate into existing business processes.
Get more bang for you bucks
As suggested by Krishna, the move towards using smaller language models represents a cost saving—something the accountants will be fully on board with! But from an efficiency and process optimisation perspective, this cost reduction represents an opportunity to solve many more problems.
Once the initial use-case has proven its value then future business cases become easier to write, and with efficiency and cost-saving data to back this up, it’ll be a cold-hearted decision maker who rejects the follow-up projects.
The rise of AI Agents
It wouldn’t be a tech conference if AI Agents weren’t mentioned, and Arvind Krishna to great pleasure in announcing the launch of IBM watsonx Orchestrate. If you’ve not seen our page about IBM watsonx Orchestrate’s capabilities, you can find it here.
Krishna suggested that more than 1 billion new applications will be built in the next four years, and around a third of them will incorporate AI agents. This might set alarm bells ringing for those working in AppDev who have never built an AI Agent in their life.
However, the Orchestrate family is here to help. It is claimed that the Agent Builder tool will help you build an agent in under 5 minutes using a user-friendly no-code interface. What’s more, Orchestrate currently provides 150 pre-built agents which can seamlessly be integrated into your workflows. These agents support work functions including HR, sales, procurement and IT, and it is inevitable that more will become available.
You can see a demo of Orchestrate Agent Builder in action here.
Where next for you?
As Arvind Krishna mentioned in his keynote, there is a wealth of data within each and every organisation that could be accessed and utilised to make better business decisions and speed up processes, and AI is at the heart of enabling this.
It’s not about pushing all your secure data to an open AI tool. It’s about identifying where your efficiencies could be found and developing a solution to achieve your goals.
It’s also about remembering that you’re not on your own. You don’t have to figure it all out by yourself. There are businesses out there (such as DeeperThanBlue) who will work with you to understand where your pain points are and identify the best way to overcome them.