Outsourcing Software Development: It's Changed With AI
Shaveer Mirpuri is cofounder and CEO of Insite AI, an AI and strategy partner for large consumer brands.
Would you hire a plumber who didn’t own any tools? Should you hire a copywriter who doesn’t thoughtfully leverage generative AI?
In the fast-paced world of AI, it’s common for businesses today to partner with offshore firms that can provide talented people specializing in advanced data science, AI development and software engineering. The benefits of outsourcing include speed, expertise—and in certain regions—cost efficiency.
Outsourcing isn’t anything new. From my view, I’d estimate that U.S.-based consumer goods companies perform about 30% of software and analytics development onshore and 70% offshore. Yet with tangible advancements in AI, companies must expect more from the firms with which they outsource. Staff augmentation companies are quick to pitch the benefits of AI to large organizations but rarely use AI themselves to get their work done.
There is a new expectation in outsourcing. Staff augmentation companies must now evolve to bring custom-built AI toolkits and components for their customers' specific needs, to pump out software and AI development at a better speed, quality and cost than before.
A Need For Staff Augmentation
According to the IBM Global AI Adoption Index, the leading barrier to AI adoption inside a company is talent. More than a third of the IT executives surveyed said they are hampered by people with limited AI skill sets.
In the consumer goods industry, 7% of global product marketers are employed by CPG companies, while only 1% of data analysts or scientists are on their payrolls, according to a study by the Josh Bersin Company.
There are clear cost-savings when offshoring data scientists and data engineers. For instance, a company overseas can supply a large U.S.-based organization with an exclusive team of AI-trained engineers at $50 to $125 an hour per person, for as many months as needed. By comparison, should that same large organization employ an onshore, full-time AI-trained engineer, it could require an outlay of mid-six-figure cash compensation, stock, benefits, office space, equipment and training. That one full-time employee is effectively going to cost $175 to $500 an hour, depending on experience.
Of course, outsourcing can come with its challenges. Outsourced teams may not have a deep understanding of the business context, industry-specific requirements or strategic technical goals—which can impact the relevance and effectiveness of AI solutions. That can be solved by picking an industry-specialized staff augmentation partner, but AI has unlocked something new.
What To Look For In A Staff Augmentation Partner
Ensure a staff augmentation partner is assigning you people who will bring AI code and use that AI code to execute software development or data science tasks.
Across industries, the expectation today is for outsourcing firms to bring their own AI-powered tools that enable their developers to work better, faster and with more intelligence. These partners should have at the ready libraries, algorithms, repositories of code, components, toolkits or pre-built solutions that derive richer results for that client.
For example, in construction or engineering, companies engage offshore civil engineers for the creation of structural drawings, engineering specification documents, bill of material analysis and so much more. The outsourcing partner needs to bring customized AI to speed up the work done by those civil engineers and bring novel intellectual property specific to the sector, as opposed to generic open-source tools.
In the CPG space, a company seeking advanced data science staffing requires a partner that has industry-specific AI forecasting models and will deploy that data science staffing to use those AI models. With this understanding and toolkit in hand, the AI-powered data scientist comes to the job with more granular and accurate results.
If a consumer brand outsources a data engineer to process and analyze a mountain of Amazon sales data, they should expect that the AI-powered data engineer comes with AI code that partially automates the extraction and representation of sales data from Amazon.
At the least, firms leveraging staff augmentation partners in analytics or AI should expect that the staff they are assigned come with libraries that generate explainable results. An AI-powered engineer won’t just deliver your tasks to forecast the uptick in sales at convenience stores but will convincingly demonstrate any causality with external factors, such as gas prices.
Expectations In Outsourcing In 2024
For a consumer goods company, outsourcing data scientists to augment a brand team’s analytics can be done easily and cost-effectively. When done right, I’ve seen outsourcing engagements result in significant cost savings—more than 30% increases in project or output quality, and a more than 40% reduction in time-to-project completion.
Companies across industries are already outsourcing in many ways, including for AI-powered insights and predictive analytics. As we get deeper into the new year, it’s important to push outsourcing partners to bring more to the table. Firms relying on AI to deliver more efficient results need to make sure outsourcing partners can deliver more customized results. They need to bring custom tools to the engagement: AI toolkits, libraries, dashboards and components that populate data and deliver insights in personalized ways.
Outsource smarter by looking deeper at the AI technology your partners have built for their people to gain efficiencies, which they can pass on to you.
Source: Forbes