AnswerRocket today announced it has acquired Cognitive Spark to increase the scope of the consulting expertise it can apply to artificial intelligence (AI) projects.
While AnswerRocket has previously provided specific AI product implementation capabilities around its Max AI platform for building AI agents, the acquisition of Cognitive Spark expands the business process expertise the combined company will be able to bring to bear, says Jim Johnson, managing partner for AI solutions and consulting at AnswerRocket.
That’s critical at a time when many organizations are finding it challenging to effectively operationalize AI technologies, he added.
Most larger organizations have hundreds of experimental projects but identifying the projects that will succeed has for many of them become a significant challenge. Specifically, many organizations are discovering that generative AI technologies lend themselves better to processes that require some type of judgment to be made using the probabilistic reasoning capabilities embedded within a large language mode (LLM), he noted.
The challenge is that LLMs don’t generate the same output the same way every time. If a process is deterministic in the sense that it needs to be completed the same way every time, then an LLM might not be the best approach to automating those types of tasks, Johnson highlighted.
Most business processes, he said, are a mix of probabilistic and deterministic tasks that organizations need to better understand before applying generative AI. In addition, organizations need to ensure that LLMs are allowed to admit they don’t know the answer to a question. Otherwise, they will strive to generate a response no matter how improbably that suggested outcome might actually be, he cautioned.
Ultimately, AI technologies will drive a massive wave of business process re-engineering. Much of the focus today is using AI technologies to automate existing tasks versus reworking how those tasks might be completed using, for example, AI agents functioning with the confines of a set of guardrails supervised by humans, Johnson noted. “A human is always going to need to be in the loop,” he told.
Finally, organizations need to realize the extent to which AI models need to be monitored and observed to ensure that as they are exposed to new data, the scope of the answer being provided doesn’t start to drift away from the original intent.
The pace at which organizations are successfully adopting AI technologies is naturally going to vary widely, but it’s clear a lot more effort is going to be required than merely exposing data to an LLM. The issue is that while many organizations have a fair amount of internal IT expertise, those teams don’t always have the experience needed to identify which projects are more likely to succeed than others.
Additionally, consultants can help identify AI capabilities that are rapidly becoming table stakes that every organization will need to have, versus, projects that are more likely to provide a sustainable competitive advantage. Regardless of which side of the ledger those projects fall, however, the one thing that is certain is that organizations that are not embracing AI technologies will soon find themselves falling way behind competitors that are developing the hands-on expertise today that will be needed to succeed tomorrow.
