The Wall Street Journal recently held its Future of Everything Festival in New York remarks during one of its key sessions from Arvind Krishna, IBM’s senior vice president of cloud and cognitive software made a headline afterward. “Data Challenges Are Halting AI Projects, IBM Executive Says,” wrote WSJ reporter Jared Council in the publication’s CIO Journal.
Here at Gravyty, we found this assessment of the state of artificial intelligence troubling. Here’s one of the top companies in the world, IBM, admitting that AI progress is stymied because customers aren’t prepared for implementations and data mapping. These are issues that plague IBM and the majority of AI vendors out there, but that Gravyty – a much smaller organization – has overcome.
We overcame this issue by making it our own. Unlike large technology companies, we do not have the luxury of churning and burning through customers. Rather than accepting failure or placing the blame on typical issues with preparing data in ways that make our AI products useful, we fight hard against the industry standard that Krishna cited in his session: “In the world of IT in general, about 50 percent of projects run either late, over budget or get halted.” If a technology vendor is not proactively working to map a customer’s data to make a use case or application successful, what are they getting paid for?
The Gravyty team prides itself on an AI implementation process that lasts only a few weeks and gets a customer fully live so we can collectively start measuring toward success. We accomplish this by making any data challenges that might impede progress our team’s mission to overcome. We don’t accept the status quo that Krishna references.
Are there best practices for successful AI implementations to avoid halting a project?
Absolutely. Gravyty has found that, like our team, customers want to change the status quo of halted, mishandled, and failed implementations. We recommend the following if you are considering an AI solution:
- Identify a vendor that you consider a partner. Make sure that they are committed to your long-term success. From establishing a specific use-case to acting as an advocate to ensure that you, as a customer, get what you expect from the technology, partnerships are less transactional and focus on the right fit for both parties.
- Work with AI partners who are clear about timelines and proactively come with plans to achieve tasks on-time. The majority of problems that IBM’s Krishna identified above centered around timing. Great partners not only work stick to timelines around implementation but also work with you to shorten time to value (TTV).
- Know how your team maps with the team of your AI partner. Just as members of your team have different roles and responsibilities, that’s also true for your AI vendor. It’s important to map these roles and responsibilities out for everyone involved in the project. Who from your team does granular work like mapping data with your AI partner, and who is their counterpart? They should be on a first-name basis. There should also be similar structures for every element involved in the process, from data security to day-to-day user experience, to regular high-level reporting.
- Set up a process for exchanging feedback. We all know that communications come to define vendor-customer relationships. Sharing feedback is a vitally important element, and, whether through formalized feedback looks or one-off clarifications and questions, all parties involved in the relationship should be empowered to share feedback. Further, make sure that your AI vendor is open to attaching suggested timelines to any feedback which requires action to ensure that your concerns are properly prioritized.
At Gravyty, we believe that there’s no reason for data challenges to halt a customer’s AI progress. If you’re interested to learn more about engaging internal stakeholders around AI, see our recent white paper here.