When most companies were still learning how to log on, Publicis Sapient was helping enterprises step into the web age. Three decades later, the company is once again working towards leading another transition, this time from digital to an AI-driven transformation.
“Back in the early 1990s, the company was set up to help enterprises adopt the internet,” said Rakesh Ravuri, CTO & SVP – engineering at Publicis Sapient, in an exclusive conversation with AIM.
Now, the company is attempting to rewrite that legacy with a $100 million self-disruption initiative, one designed to use AI not just as a tool, but as a co-creator across industries.
From Media to Machine Intelligence
When ChatGPT ignited the generative AI race, Publicis saw the writing on the wall. The group had already acquired Epsilon, a major data broker, and Lotame, giving it access to over 90% of third-party user data globally. With such data depth and decades of industry know-how, the company moved quickly to create platforms that bring AI to the core of business transformation.
“Two years back, our group CEO announced that we would invest $100 million over three years to disrupt ourselves,” said Ravuri. “We created platforms like CoreAI for marketing transformation, and the Sapient Slingshot, which reimagines the software development lifecycle using AI.”
While CoreAI leverages Epsilon’s massive data engine to personalise campaigns, Slingshot is where the engineering revolution begins.
An AI-powered platform designed to transform the software development lifecycle for enterprises, Slingshot connects every stage of development, from strategy and design to deployment, through a unified context graph that synchronises information across teams, tools, and repositories. The platform includes intelligent assistants for developers, product managers, and designers, all tapping into shared contextual data.
This approach allows enterprises to automate coding, documentation, and testing processes while maintaining control, accuracy, and alignment across large-scale development environments.
“The problem with tools like Copilot or Cursor is they operate only at a workspace level,” said Ravuri. “Our context graph connects the frontend, backend, and every repository in real time. So when someone makes a change in one system, every related system knows it instantly.”
That connected fabric extends beyond development. Platforms like Bodhi, an ML Ops environment, and Sustain.ai, an AI operations layer, close the loop by adding predictive maintenance, recommendation, and optimisation capabilities. Together, they form what Ravuri calls Publicis Sapient’s “people plus products” approach to transformation.
Modernising the Past to Build the Future
Publicis Sapient’s AI journey involved not only building new things, but also modernising the old. “When we started, companies told us they wanted to use AI, but their systems were legacy,’” says Ravuri, adding, “We found organisations still running on 60-year-old code.”
That insight led to one of the company’s fastest-growing businesses: legacy modernisation. Publicis Sapient converts COBOL code into modern languages like Java, while also enhancing security, accessibility, and compliance requirements in the process.
Unlike brute-force code-to-code conversion tools, the system takes a multi-step approach, from code-to-spec, spec-to-design, and design enhancement, to spec-to-code. Each phase uses specialised AI agents that reference the same context graph to ensure continuity and correctness.
“We’ve completed projects for global banks in Australia and the UK, healthcare providers, and even government agencies in Saudi Arabia and Dubai,” said Ravuri. “The reason clients prefer us is that our agents don’t just translate — they transform.”
The company’s on-premises capability has also made it attractive to clients in sensitive sectors. “We deployed the entire system inside the Saudi government’s data centre,” he said. “They wanted their own Arabic model, and we were able to integrate it easily into our platform.”
Redefining the Engineer’s Role
If AI is rewriting the code, the human role is being rewritten too. Ravuri believes the future of engineering lies not in typing faster, but in thinking deeper.
“Everyone will have a team of agents working for them, and divide the tasks between agents and humans,” he said, indicating how an engineer’s role will be about guiding the agent.
He sees this shift as both an opportunity and a challenge. Reviewing AI-generated output requires a strong grasp of computer science fundamentals, not just frameworks or syntax. Ravuri warned against blindly accepting AI-generated content, stressing the necessity of understanding algorithms, design patterns, and distributed systems to effectively review such outputs.
To him, the rise of AI in coding mirrors an earlier phase of automation, but with a crucial difference: it highlights that AI operator roles may not be high-value jobs.
The higher-value roles, he insists, will belong to those who can review, guide, and design AI systems, the new managers of machine intelligence.
The Human in the Loop
For all its automation, Publicis Sapient’s vision of AI still keeps a person in the loop. Their virtual teammate operates in the background, identifying issues and even generating branches with potential solutions. However, the critical final steps of review and merging still require human expertise.
It’s a pragmatic view of a future where AI isn’t replacing engineers but reshaping how they work. Ravuri called it “AI stewardship”, a skill that will soon define the best developers in every team.
“AI can generate ten times more output than a human,” he said. This makes reviewing significantly more critical. Consequently, a strong foundation in computer science, previously considered optional, is now absolutely essential.
From its MIT roots to its AI-led platforms, Publicis Sapient’s story has always revolved around one word: context. Whether transforming banks, ads, or codebases, it has sought not just to digitise work but to make it intelligent, connected, and human-guided.
As Ravuri puts it, “AI is only as good as the context.”
The post How Publicis Sapient Is Rebuilding Software Development Around AI Context appeared first on Analytics India Magazine.
