Our Report TL;DRs serve up a quick look at recently-published reports worth paying attention to, and sharing with your team. Important note: our editorial team and contributors select these. The firms producing these reports don’t pay to be featured.
• • •
Report Title
Bain Technology Report 2024: Technology meets the moment as AI delivers results (PDF version here) — Bain
Published
September 25, 2024
Most Useful For
C-suite, CTOs and CIOs, software development leaders, and business unit innovators — especially those tasked with implementing generative AI.
Data Sources
Primary data comes from polls conducted by Bain in 2024, including an IT Decisions Makers Survey conducted in January, with 151 respondents, and an IT Workload Survey conducted in May, with 283 respondents. Predictions are based upon Bain analysis of analyst reports and conversations with market participants, as well as commercial data sources like IDC, Gartner, and Bloomberg.
Software development especially has seen significant speedups in creating code and documentation, and customer service teams can cut response times up to 35 percent.
Key Points
- Businesses are continuing to invest in generative AI, with the number of large companies investing more than $100 million more than doubling in a year. And early adopters are reporting that it’s paying off, with performance gains up to 20 percent of earnings after 18 to 36 months.
- Top AI use cases for productivity gains include software and product development, customer support, sales and marketing, and back office operations. Software development especially has seen significant speedups in creating code and documentation, and customer service teams can cut response times up to 35 percent.
- Businesses shouldn’t just plug AI into existing processes and computer systems and hope for the best. By optimizing business processes and getting data well-organized in modern database systems with a single source of truth, and being deliberate about prompt engineering and AI application design, they can see better results.
- Generative AI can save 10 percent to 15 percent of software engineering time in particular. Companies should make sure they’re using efficient engineering processes like continuous integration and deployment, and good metrics to track developer efficiency, to make sure this newfound time is used efficiently and that software can ship as quickly as it’s created.
- Increased demand for AI may lead to shortages of graphics processing units (GPUs) and boost demand for labor and electricity to build and operate data centers. Business will want to move away from just-in-time inventory strategies, diversify supply chains, and potentially secure long-term supply agreements for AI-related hardware and services.
One Great Chart
How to Apply These Insights
Choose Where to Use AI: You can explore possible AI use cases through pilot programs, but certain areas like customer service and software development often prove more fruitful than others. Invest in deploying and refining AI where it can lead to the biggest productivity increases, based on industry data and what you see in your company.
Plan Ahead: Organizing data for AI to ingest and optimizing business processes—including software development practices—can help generative AI deliver more for your business. Think about which AI tools and vendors to use and crafting proper prompts for generative AI, and plan for any disruptions in supply in the rapidly changing field.
Measure, Measure, Measure: Make sure you have metrics for tracking how AI transforms factors like productivity, customer satisfaction, and employee morale. Keep an eye on the numbers to ensure your AI investment is delivering and that freed up resources like employee time are used efficiently.
Questions to Discuss with Your Team
- What metrics are we tracking as we experiment with and deploy generative AI?
- How can we prepare our data and processes to make AI more useful for us?
- Do we have the internal know-how we need to work most efficiently with AI systems or do we need to hire or bring in external help?
- In what ways can we help our software developers work with AI most efficiently?
- How prepared are we for hardware supply chain disruptions, or changes in the availability or cost of AI cloud services?
