This is the final article of a 3-part series, "The Case to Embrace AI." This series examines AI usage among professional services firms, and more specifically, how some consulting firms, as focused as they may be on guiding clients on their AI journeys, could be overlooking important opportunities to put AI to work within their own companies. The series will outline some of the promising use cases for AI in a professional services firm, the benefits these use cases can provide, and the foundational elements firms need to put in place to reap those benefits.
Succeeding With AI: A Practical Guide for Consulting Firms. Mitigating risk, targeting the right internal use cases and starting small are key to unlocking AI's power. Contributed by Carlos Sanchez.
The consulting business and the broader world of professional services aren't shy about opening their wallets for AI. By 2027, according to a 2023 survey from IDC, worldwide spending on artificial intelligence capabilities is projected to exceed $423 billion, with professional services and three other industries expected to be most aggressive, accounting for 60% of global IT spending on AI.
Clearly, professional services firms see a solid business case for investing in business AI (including generative AI, or genAI). Indeed, as I discussed in parts 1 and 2 of this series, the case for consulting firms to embrace AI within their own businesses (as well as in the services they offer clients) is highly compelling, giving them the means to automate processes, better support clients, boost productivity, improve resource management, and as the foundation for new business models and revenue streams.
As with any technology or digital system, AI comes with risks that organizations planning to or already using the technology would be wise to address in order to protect their investments. A lack of explainability with an AI model, bias in the data or algorithm that skews outcomes, production of suspect information, displacement of human workers — these are among the risks that a firm must consider when wading into the AI waters. And recently we were reminded that cyberattack is a real threat with AI, too.
However, these risks shouldn't discourage consulting organizations from exploring AI use cases inside their business, as potentially impactful as those use cases can be. Rather, they would be wise to approach AI as they would any new technology: with an open mind and a well-thought-out, tactical approach in how they invest in and deploy it, conducting a thorough cost-benefit analysis and risk assessment as part of the process.
As your firm embarks on the AI journey, it's important to realize that not all business issues warrant an AI solution — a traditional software solution might suffice. Where AI is the best solution, be sure to focus on identifying use cases that are technically feasible, map to a well-defined business need or problem, align with your business goals, and project to add enough value to the business to justify the investment. In the IDC survey, North American corporate IT buyers reported that AI use cases involving the automation of IT tasks provided the highest return on investment, followed by use cases that yield product and service innovation.
After identifying AI use cases worth pursuing, one of your next big to-do's is to find a technology partner or partners that can provide you with what you need to build out those use cases. If yours is a deeply resourced firm with a large in-house stable of AI-savvy developers (few consulting firms fit this profile), then maybe all you need to source externally is an open-source AI framework with a large language model. Many firms, however, likely will need a multifaceted set of AI capabilities that includes a variety of models to test, along with storage and computing/processing capacity, data resources and other tools, all accessed within some kind of development platform or environment. Some may prefer a plug-and-play, off-the-shelf type of industry-specific solution that maps to professional services-related use cases. Here's where thorough due diligence in evaluating what you need, as well as the provider(s) of the AI platform, models or off-the-shelf AI solution — the security measures they have in place, their commitment to ethical use of AI, etc. — is a must, keeping in mind that you may end up relying on more than one technology vendor in your AI journey.
In evaluating use cases and specific AI products, it's also important to consider how readily the use case and product can scale and be applied elsewhere in your business. The more scalable and adaptable an AI solution is (without sacrificing the quality of outcomes), the lower the marginal cost associated with your AI investment will be.
Also as part of the process, be sure to keep AI ethics/governance, as well as staff AI training needs, front of mind, as these likely will be major factors in the overall success of your AI endeavors. Having a clear, comprehensive AI ethics and governance policy, one that people throughout the organization understand and follow, is a must. For ideas about what to include in such a policy, try a Google search to see what other companies are doing.
When your cost-benefit and risk analyses are complete and you've laid the fundamental groundwork (including assessing the state of your data and taking the necessary steps to ensure it's fresh, comprehensive, trustworthy and readily accessible to train your AI models, as I discussed in Part 1), it's time to put AI to work inside your business. One piece of advice here: As ambitious as your ultimate plan for using AI might be, it's wise to start with a narrow proof-of-concept project to test the waters. Integrating business AI into a firm carries substantial expense and risk, so you want to be confident the technology can deliver the outcomes you expect, and do so cost-effectively, with a manageable level of risk.
Say, for example, the use case involves a workforce management solution embedded with a virtual assistant to help match in-house resources to specific projects. In the proof of concept (POC), you'll want to evaluate the quality of the user experience and of the recommendations the virtual assistant provides based on the data the solution's underlying model used, and how it was prompted. You then can refine the model, along with the prompts and the data it uses (with help from your technology partner, which could be the provider of the AI solution and/or a third-party implementation specialist). Keep iterating until you get the outcomes you expect for the use case, or until it's evident the AI solution isn't the right fit for this use case. You also want to be sure the AI solution, its underlying model, and the output it produces are readily explainable, not only to mitigate potential legal, security and compliance issues, but also to inspire trust in the technology and enable auditability to ensure model performance doesn't drift or degrade over time.
Ultimately, these POCs can lead in surprising new directions, revealing unexpected possibilities for iterating an AI model to add value in other use cases within your business. So be alert to opportunities to apply AI capabilities in areas you initially may not have contemplated using them.
"By deploying applications that help enhance productivity and personalize processes," the Deloitte AI Institute said in September 2023, "organizations can use Generative AI to accelerate the pace of their business, evolving an experimental investment into an established value driver."
The sooner your firm starts responsibly experimenting with business AI, the sooner it can start stepping on the accelerator.
Carlos Sanchez, Platinum Business Principal, SAP Carlos Sanchez has been a business and technical consultant for 25 years with multiple consulting organizations. He is currently a solution expert in the SAP Professional Services Global Industry Business Unit.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.