July 11, 2023
Just over six months after the's debut of ChatGPT, software companies worldwide are using the underlying technology to speed up development and enhance their customers’ products.
If your social media feeds are anything like mine, they’re full of posts and adverts outlining some intriguing new use for generative AI systems like GPT-4, Bard, or Midjourney.
“There’s an app for that” was the common refrain at the start of the 2010s as the rise of the smartphone spurred a flurry of app development. Now generative AI is being touted as the tool to transform everything from recipe recommendations to discovering new drugs.
It’s a fast-changing sector with a new wave of advances coming every week. But a survey of Accelerance partners around the world suggests it's more than the latest tech fad.
Executives from big and small companies, seasoned technology professionals who have witnessed the debut of successive waves of technology, have outlined a wide range of applications of generative AI that they’ve already deployed internally and shipped in products.
“It is going to transform all business as it keeps evolving,” one software lead told us.
“It has crossed the hype cycle already,” was how another put it.
Old school AI
Many of our partners were veteran users of artificial intelligence and machine learning long before large language models and generative pre-trained transformers came to dominate the field of AI.
One survey respondent, a South American software developer, had helped a company in Spain search for machinery on city streets in need of repair.
“The algorithm learns about the issues of these street machines, so they solve those problems each time faster,” he told us.
Accelerance partners have long deployed AI-powered computer vision systems.
“We have up to 10 ongoing projects, such as a computer vision solution for label detection in warehouses for a global Fortune 100 engineering company,” a manager at a large development company with staff spread throughout Europe told us.
“We developed an AI-driven payment approving decision engine for a UK digital bank and predictive maintenance for an in-flight connectivity provider.”
Numerous respondents noted their widespread use of AI in generating business analytics, data insights, business process automation, and robotic process automation.
Beyond the concept stage
So far, so normal. But they’ve also spent the first half of the year experimenting with generative AI and shipping software that employs it.
ChatGPT underpinned a voice-based chatbot one of our partners with operations in South Asia and North America has created.
“The solution enables real-time interactions with health-related queries for pets and supports mobile as well as fixed-line interfaces,” the partner told us.
Another large partner was helping its customers streamline their quality assurance processes by deploying interns to help them engineer the right prompts for applying generative AI.
A Bangladesh-based development house is using large language models (LLMs) to break out specific information from a customers’ financial reports to analyze their employee salary structure. It has also created a chatbot that integrates with WhatsApp to answer employee questions based on a company’s internal information.
Another solution that generated business analytics and automated reports for meaningful insights, enabling them to leverage data in precise decision-making using statistical methods.
Numerous developers are integrating OpenAI’s GPT-3.5 and GPT-4 services via application programming interfaces they are employing for their own services and those of their customers.
How AI is augmenting software development
If generative AI is already making its way into customers’ products, it’s also being used in the software development process itself.
“Our internal R&D team has been evaluating how we can apply ChatGPT and other generative AI tools in our projects,” a Malta-headquartered partner told us.
“The early estimates showed that they can save around 15% of the time for development and costs.”
Development teams have widely adopted Github Copilot, autocompleting snippets of code in development environments like Visual Studio, Neovim, and JetBrains.
Our partners are also moving rapidly to upskill their developers and consultants in the largest generative AI toolsets.
“Besides the team members that are already working on client's engagements, nowadays we have a community of practice that has around 1600 volunteers working testing different tools, experimenting with proof of concepts (PoCs) and minimal viable products (MVPs),” one publicly-listed multinational software partner told us.
Another summed up the position of many partners when it comes to generative AI:
“Rest assured, we are actively working towards meeting the demand for generative AI and delivering exceptional solutions to our clients.”
True potential, with a side of hype
When we asked our partners whether they put the current interest in generative AI down to the usual marketing hype around new tech, or a response to the genuinely transformative nature of it, most leaned towards the latter.
“It’s transformative and strategic,” one respondent noted.
“Definitely the real deal,” wrote another.
Still, some view the current frenzy of interest in the technology as grounds for caution.
“While the technology offers great potential, it is crucial to understand its limitations and ethical implications. As responsible practitioners, we must utilize generative AI tools thoughtfully, ensuring they are used for positive and beneficial purposes,” a partner with developers based all over South America told us.
“The current hype is excessive and will generate a large amount of wasted efforts from many companies that are not stopping to analyze the potential impact of adopting these tools,” a partner with operations in Panama and Spain responded.
So how do you avoid wasted effort and maximize your return on investment in developing and deploying generative AI in your products and services?
Based on the experience of our consultants and the wealth of knowledge from our global network of partners, here are eight things you should consider before investing in building generative AI systems:
- Define clear objectives: It's crucial to establish specific goals and objectives for implementing generative AI systems. Determine what problems or opportunities the AI system will address and how it will benefit the business.
- Data availability and quality: Generative AI systems require substantial amounts of training data to produce meaningful and reliable results. Assess the availability and quality of the data needed to train the AI system. It should be representative, diverse, and of sufficient volume to ensure accurate modeling.
- Expertise and resources: Evaluate whether your business possesses the necessary skills and resources to develop and maintain such systems. If not, consider partnering with external experts or investing in training your existing workforce.
- Ethical considerations: Generative AI systems have the potential to generate biased, unfair, or malicious content if not carefully designed and monitored. Consider the ethical implications of the AI system, such as privacy, security, transparency, and potential social impact. Develop guidelines and mechanisms to ensure ethical behavior and mitigate potential risks.
- Regulatory and legal compliance: Be aware of relevant laws, regulations, and industry standards that govern the use of AI systems. Ensure that your AI system complies with applicable data protection, intellectual property, and privacy regulations.
- Scalability and integration: Consider how the generative AI system will integrate into your existing infrastructure and workflows. Assess its scalability to accommodate future growth and increased demand. Determine if the system can be easily integrated with other tools, software, or platforms used by your business.
- Return on investment (ROI): Conduct a thorough cost-benefit analysis to evaluate the potential return on investment. Assess the expected benefits the AI system will bring, such as increased productivity, efficiency, cost savings, or improved customer experience. Compare these benefits with the costs associated with development, implementation, and maintenance of the AI system.
- Continuous improvement and monitoring: Generative AI systems require ongoing monitoring, maintenance, and improvement. Consider how you will collect feedback, evaluate system performance, and iteratively enhance the AI system over time. Plan for regular updates and adaptations to keep up with changing business needs and technological advancements.
By carefully considering these factors, you can make informed decisions about investing in generative AI systems and maximize your chances of successful implementation and utilization.
Get in touch with Accelerance’s experts to find out how our global network of partners can help you cut through the hype and take advantage of generative AI.
