We have covered several topics related to process and knowledge in previous blog posts on how to become ”AI Ready”. Of course there will have to be some technology involved also when you work with digital innovation and AI. In our ”AI Ready Model” we have decided to describe the technology required to fulfil your AI requirements as a capability.
The development related to AI is rapidly evolving and there are already a lot of various AI models and technology available on the market today. Many specific applications (like ERP systems, collaboration suites, office suites etc) already incorporate various AI models and you can use AI assistants to help you translate from different languages, write your CV, solve your math homework, help with your coding or many other use cases.
The question is which of these models you can trust and what happens with the data that you willingly share with them? Maybe this is harmless for the use cases described above that only concerns your own private data, where you decide what you do with that and which models to use and trust.
However, when at your workplace you represent an organisation that is required to comply with legislation, that might not want to share large chunks of business data with anyone, is heavily reliant on the response of the AI model or tons of other potential problems you that you might need to address.
How to avoid mistakes
When implementing AI to address business opportunities or challenges, there are many of these considerations you need to do before putting something in production and in the hands of your users. This is why we have decided to list Technology as a capability in the model. We don’t think that these are considerations that should be put on individuals or projects that wants to implement AI tech in the business. There should be a centrally managed capability available that provides guidance, advice and verified AI technology to the business initiatives.
This way you can avoid that initiatives make mistakes with using non tested, unverified tech that ”hallucinates” (a well-known phenomena with some AI models) or shares data in an undesired way. The capability of tech in the ”AI Ready Model” should contain and provide AI solutions that has been tested and verified for use in the organisation.
They should be accompanied by instructions, recommendations and other useful information for the initiatives that wants to implement the technology. This way the organisation can build collective knowledge and competence on AI tech, rather than limiting it to individuals or specific projects.
Evaluate for success
The AI tech recommendations should include documentation on how the models/applications are deployed and where the data put in really ends-up. Today most AI models run on some kind of cloud infrastructure, often deployed on one of the global cloud platforms. It is often hard for the ordinary user to understand if this data is stored somewhere after usage or if it is even used to train the model for the next question (this way the AI model can potentially relay knowledge that has been gained through your business data).
These are things that should be evaluated and documented within the ”AI Technology” capability to make it easier for users to pick a model or product ”off the shelf” when initiating an AI initiative. Fully functioning in the intended way in our model, the ”Technology” capability of the AI model should be viewed upon as a catalogue or repository of tested and verified AI tech, that can be utilised in the organisation for business use cases.
For each alternative there should be a description and documentation available to make it easy to implement for each unique case. To achieve this, you will need to assign resources with AI models and technology knowledge to evaluate which technology to make available for your organisation. New technology can of course be evaluated and added over time.
A review process for re-evaluation of previously evaluated technology should also be implemented, as conditions can vary over time.
More information
Rightly managed Technology will speed up your AI initiatives and provide them with Quality Assured and verified AI models and tools to generate business value from AI. This is why ”AI Technology” is a cornerstone and an important capability of the ”AI Ready Model”.
