The promise of automation is easy to sell from a demo. The harder part is checking what sits behind the interface — and that gap is now costing Indian marketers real money.
A drive-through AI that still needed people
The US Securities and Exchange Commission acted against Presto Automation in January 2025 after finding its AI voice technology did not work as claimed. The original version, deployed at restaurant drive-throughs in September 2022, required human intervention for every single order, according to the SEC. A later version piloted from June 2023 still needed people to enter roughly 70% of orders through at least December that year.
The company had told investors the technology eliminated human order-taking. It settled the matter without admitting or denying the findings, and no civil penalty was imposed.
That case is American, but the question it raises belongs in every Indian boardroom approving an AI purchase: what exactly is the software doing?
India’s AI spend is racing ahead of readiness
Indian appetite for generative AI is real. An EY-CII study published in November 2025, covering 200 organisations, found 47% had multiple generative AI applications in production, while just 10% were scaling them across the business. Speed of deployment influenced buy-versus-build decisions for 91% of respondents.
That urgency creates exactly the kind of pressure where inflated claims thrive. The buyer wants a quick operational fix; the seller has a reason to make implementation sound simple. Data, integration, supervision and post-demo work become expensive afterthoughts.
Global figures show the same strain. Gartner’s 2026 CMO Spend Survey, based on 401 respondents from North America, the UK and Europe, found marketers allocated an average 15.3% of their budgets to AI initiatives. Only 30% reported mature or fully developed AI readiness.
Gartner’s Ewan McIntyre, VP Analyst and Chief of Research for marketing, warned: “AI maturity is beginning to separate marketing leaders from laggards.”
A separate BCG survey of 300 global CMOs found 96% reported significant transformation through AI, but 42% were still using generative AI mainly to assist people in limited workflows. Only 32% were deploying agents across redesigned workflows. In the same survey, 43% of CMOs said their AI investment exceeded $15 million.
How AI washing reaches the invoice
AI washing means a supplier claims a capability the underlying technology cannot support. Gartner said in June 2025 that thousands of vendors were relabelling assistants, chatbots and robotic process automation as agentic AI without genuine agentic capabilities. It estimated only about 130 of those vendors were genuine, and forecast more than 40% of agentic AI projects would be cancelled by end-2027 because of cost, unclear value or weak risk controls.
Anushree Verma, Senior Director Analyst at Gartner, said: “Many use cases positioned as agentic today don’t require agentic implementations.”
The SEC has also penalised misrepresentation. In March 2024, investment advisers Delphia and Global Predictions agreed to combined civil penalties of $400,000 over misleading AI claims, without admitting or denying the findings.
What to check before signing
For Indian brands, the useful dividing line is evidence. A polished demo on familiar English queries proves little about real customer language, accent, incomplete records or integration with existing systems.
Before buying an AI marketing or support tool, marketers should build the following into evaluation:
- Define the promised function: what exactly must the AI complete without human handoff?
- Set acceptable error rates and escalation thresholds — and confirm they include the vendor’s own staff, not just yours.
- Ask what the supplier owns versus what it licenses from another model company; this changes pricing, continuity and accountability.
- Run a controlled comparison against the current workflow, measuring completed resolutions, not just generated responses.
Hindustan Unilever’s descriptions show what clearer AI claims look like. Chairman Nitin Paranjpe described AI-powered personalised advertisements inside the company’s Shikhar retailer platform, alongside its ordering and CRM tools. Former CEO Rohit Jawa told the ISA CEO conference in February 2025: “So, a clear strategy and a robust data foundation are absolutely essential for a successful AI implementation.”
The purchase decision should follow that same discipline. Contracts can specify the function, error rates, human involvement and the evidence required before expansion. A pilot is only useful when the buyer can test the claim — and decline the next cheque.
Source: Indian Advertising Media & Marketing News – exchange4media
