We’ve forayed into a world where the question is no longer whether to use AI, it is more how to use AI to maximum advantage. Google Cloud, as ever in the spirit of making AI more accessible, has sought to create a community project in the UK designed to empower users implementing SAP RISE projects with a set of tuned and tailored AI prompts that will do most of the heavy-lifting when it comes to some of those time-consuming components. According to Matthew, “Gen AI is a participation sport”. With this community project, Google Cloud is taking the next step to making AI accessible and ubiquitous. Find out more in the points below from an interview between Matthew Buskell, SAP RISE and Enterprise AI Lead, UK & Ireland,Google Cloud and Jonakee Chandra, MD, Wellesley Information Services.
ETHOS OF A COMMUNITY PROJECT: Google Cloud’s objective with this community project is to create a catalogue of multimodal AI prompts that individually provide impactful incremental accelerations for a RISE project and in aggregate, they can significantly reduce project duration while increasing accuracy. The genesis was in discovering how much RISE projects have in common in terms of pain-points and the extent to which fine-tuned AI prompts can help to mitigate those pain-points. This project set out to create a foundation for sharing and exchanging prompts by bringing together RISE partners to learn from each other.
RISE CHALLENGES: Typically, RISE projects are large and therefore complex, requiring navigating potential points of friction. SAP is investing heavily in making RISE implementations smoother and partners are equally as committed to that cause. This community project is targeted at a set of addressable problems – based on the challenges typically faced by the user community such as user documentation, regression testing etc. The success of a community project rests heavily on how well the community uses it and what they choose to derive from it. What Google Cloud is facilitating is enhanced user-generated prompts that interact with an AI model and provide enriched, relevant contextual information. This effectively draws from the wisdom of the masses by allowing such RISE team members to contribute and build up a bank of meaningful prompts.
BENEFITS: While savings are subjective, this foundation is flush with promise and potential. What this does is open the doors to a new regime of working with tried and tested AI prompts that save time, effort and energy while increasing speed and reducing risk.
LEVERAGING AI CAPABILITIES: Google Cloud’s AI-powered Gemini solution analyses videos of legacy SAP ECC and upgraded S4 systems, pinpointing differences in interfaces, processes, and functionalities. It then automatically generates comprehensive end-user documentation, including change management guides and training materials, accelerating user adoption and minimizing disruption during SAP S4 transitions. Additionally, it is endowed with a colossally large context window and while it could be measured in the number of tokens consumed, the more relatable qualitative aspect is that it is built on a vast foundation of input knowledge that encompasses everything from relevant publications to SAP’s Activate methodology, further augmented with entire books and videos. Plus there is ongoing enrichment from user inputs providing business context, user roles and real-life nuances associated with them.
BEST PRACTICES: Google Cloud has made a concerted effort to ensure that best practices are baked into the strong base of built-in knowledge that the prompts are trained on, including Activate. As with anything AI, it is dynamic and partners continue to hone the knowledge with industry-specific inputs.
PREREQUISITES: Users will need access to Gemini via Vertex AI, for the optimal results. This provides users with access to our most powerful models to train users up on tooling and technology, along with additional security controls that further protect data. The private version of this technology would be appropriate if customers were to integrate the prompts, for example, with their own business applications.
BEYOND RISE: While this community project was created with RISE in mind, it can be adapted to other types of IT projects, which tend to be characterised by a swathe of similarities. The danger with making it too generic, is that it can then become imprecise. A shared platform means that a lot can be learned within and outside the community.
COLLABORATION: Conversations are ongoing with a growing number of partners, who will start to play a bigger role. Google Cloud is very open-source and known for its sharing mentality.
CUSTOMERS: There are customer committees where participants have reviewed use cases and demos and provided feedback; some have suggested additional use cases. All of this has been incorporated into the model, which will continue to mature.
SAP’S ROLE: The ultimate landing zone for this will be the Activate methodology. In addition, These prompts are largely Gemini-specific to optimize the breadth and depth of context, including Gemini’s large context window.
MILESTONES There are three main phases:
- Gathering feedback through early customer engagement, where participants weigh in on the merits of the output being produced. E.g., savings could be immense for customers based in the FDA-regulated life sciences industry, where, ‘before’ and ‘after’ videos loaded into Gemini could produce the necessary gap analysis documentation and facilitate compliance with regulatory requirements.
- Next would be a mechanism to exchange prompts within the user community through a comprehensive and accessible catalogue.
- The third phase would be to upscale the catalogue and sharing mechanism so that they can be incorporated into partners’ respective migration methodologies.
OBSTACLES: It will need to be determined which prompts need to be private and the aim will be to build consensus around this. According to Matthew, AI has entered its third evolutionary phase, where it’s able to reason and this reasoning ability did not exist before in the earlier generations, so customers will need to stay on top of this fast-evolving technology.
THE FUTURE: Scalability of business processes is an important byproduct of AI, which is catalysed through wider adoption. In some cases, AI will be able to automate time-consuming tasks that employees to refocus their time on more impactful work.
An example from procurement, where businesses prioritise areas with the greatest impact, meaning that, while the biggest suppliers receive a fair amount of due diligence, the long tail of smaller suppliers often escape the same level of scrutiny. AI would afford businesses the scale to apply the same degree of rigour to every vendor regardless of size. In the context of RISE, AI is being applied to improve efficiency in a project framework that is already characterised by structure and definition.
SIGNING UP: Users can sign up through the survey, answer a small set of questions and scan the QR code below to join the community. However, this comes with a disclaimer, which is that it is not a frozen-in-state set of prompts and that this space is constantly evolving.
What do participants need to bring to the table?
- Some examples
- A story or a problem to solve
- Anonymised data
- Willingness to work with Google and the rest of the community in the spirit of a community project.
Link to sign up at https://forms.gle/fT5Zt9gyVcQgqVAF9