In an interaction with Finance Outlook India, Rakesh Raghuvanshi, Founder & CEO of Sekel Tech, highlights how AI and omnichannel commerce are reshaping revenue forecasting, investment strategies, and long-term ROI frameworks. Drawing on his expertise in omnichannel commerce, demand generation, geotask management, and data-driven marketing, he emphasizes that traditional channel-based financial models are becoming obsolete, with businesses needing to shift towards journey-based attribution to capture real demand. Raghuvanshi further notes that AI-led platforms integrating customer data, discovery, and performance marketing are key to improving supply chain transparency, enhancing customer engagement, and delivering measurable business outcomes, while enabling scalable digital transformation.
With over 70% of consumers engaging across multiple channels, how is omnichannel reshaping revenue forecasting models?
If your revenue forecasting method still sees channels as individual, distinct inputs, it is an accounting exercise, not a forecasting model. That is the answer most CFOs do not want to hear.
Once a customer interacts across multiple channels prior to a purchase, the linear attribution model that most financial models rely on collapses. Take, for example, the situation we observe on our OmniLocal platform in the manufacturer-dealer networks in India: A contractor in Coimbatore performs a Google search for a grinder, discovers one of the authorised dealers, views the Google reviews, calls the dealer, and walks in the next day. In this scenario, there are three touchpoints over two days. Most financial models will say either the Google search or the in-store purchase and will miss the connected journey. Revenue forecasts based on such flawed attribution models will consistently miss digital discovery's impact on driving offline transactions.
Thus, the omnichannel retail ecosystem requires CFOs and finance teams to transition from channel-based revenue models to journey-based models. The focus should shift from “what revenue did the e-commerce channel generate last quarter?” to “what revenue does a validated, high-intent customer journey across all channels generate?” Within a month of helping one of India’s leading electronics retailers align Google Business Profile (GBP) listings with their warehouse inventory, we saw a 70% increase in local listing impressions and a measurable increase in in-store traffic. These metrics had not been present in the former forecasting model because it had not connected an online discovery event with an offline conversion.
There are meaningful structural implications for finance teams. For omnichannel revenue forecasting to be accurate, there must be unified data systems across marketing, field sales, and distribution (i.e., a closed-loop attribution system). Brands that build this infrastructure first will have a compounding forecasting advantage every quarter. Those that don't will continue to misallocate spend between channels, since their model lacks full revenue visibility.
How do you build a business case for AI investments across marketing, supply chain, and customer experience?
I have been in enough meetings to understand that ‘AI investment’ as a line item under technology budgets will lose in most scenarios to other competing technology budget priorities. The opportunity cost for AI investment is never justified until a business case is constructed, and funding becomes feasible when a case for revenue recovery or cost savings is defined with a concrete, measurable, and quantifiable basis.
The approach I take is what I term as Baseline Gap Audit. Quantifying the cost your current system is incurring due to its lack of intelligence is necessary before AI investment is justified. In Marketing: how many unclaimed or incorrect authorized dealer listings are out there, and what is the estimated demand that is diverted to a competitor or unverified seller due to that gap? In Supply Chain: how much working capital is tied up in scheme credits and debit notes that are unaccounted for - in a typical 40-distributor network, this is as much as ₹4–20 crore that is frozen at any point in time. In Customer Experience: what is the value of the lifetime loss due to a gap in warranty registration capture at the point of sale and how many of them are there?
The case for investment in AI becomes self-evident and justified once the baseline loss has been determined because AI is the most cost effective way to close that gap at scale. An AI-powered management engine that monitors, corrects, and optimizes hundreds of dealer listings at the same time costs less than the demand loss that the engine will prevent. At scale, a machine learning model that instantaneously classifies scheme credits eliminates a compliance risk exposure that a human team is unable to manage reliably.
The ROI framework consists of three levels: 1) Defensive ROI. This is the revenue and working capital recovered from closing known gaps and is measurable within 90 days. 2) Competitive ROI. This is the market share gained through superior channel intelligence, and is measurable within 12 months. 3) Structural ROI. This is the compounding advantage of a data infrastructure that supports quicker, more precise decision-making and is measurable within three years. First tier ROI is typically the only tier that most CFOs sign off on, while the remaining two take on a longer-term upside.
Also Read: The Rise of Parallel Careers: How Gen Z Is Redefining Work in India
What are the biggest financial risks associated with AI adoption in omnichannel commerce — model bias, compliance, cybersecurity?
Among the many underestimated risks, model bias and cybersecurity risks are the most common. The concern that most frequently erases all positive returns from AI is what I refer to as Garbage-In Governance failures. This occurs when AI models are trained using flawed data and then systematize these flaws on a massive scale.
Take, for instance, a brand using an AI-driven demand forecasting model that is trained on sell-in data, that is, data on stock pushed to distributors, rather than sell-out data, which captures purchases made by end users. This model will predict demand based on the behavior of the distributors rather than that of the consumers. The behavior of consumers will not be a determining factor in the prediction. In a lot of distribution systems, sell-in and sell-out data can be out of alignment for extended periods of time. An AI-driven system will invariably optimize for the ‘wrong’ goal and as the AI system does this at scale, it is guaranteed that an organization will be willing to trust it to continue the misallocation until it becomes painfully obvious.
In the context of GST frameworks, compliance risk is perhaps the greatest concern. An AI-driven scheme management system that automatically categorizes and issues credits may, without the aid of a human, generate enormous compliance risk in a short time. The greatest risk isn’t the AI failing, it’s the risk of the AI succeeding, at the wrong thing, in a fully automated fashion and on a massive scale.
Most risks associated with cybersecurity and omnichannel AI data. As soon as a dataset is created, that information becomes a target. Dataset owners have many data protection obligations, and many of the first AI implementations do not prioritize those obligations. In the case of a violation, the reputational damage is often greater than any efficiency gained.
A solution exists. Get a handle on the data first. Clean, compliant, and sell data that is outsourced yields AI that is accurate, defensible, and safe to scale.
How do you evaluate the ROI of emerging channels like social commerce, conversational commerce, and AI assistants?
The primary error with new channels is using old channels' ROI metrics - cost per acquisition, return on ad spend, conversion rate - on channels that are still in the discovery and consideration phases. It makes every new channel seem unprofitable.
The model I work with is called the Demand Seeding Value model. These channels can have two types of financial contributions. One contribution is the direct transaction value they create today, and the other is the potential future demand they create in other channels. When it comes to social commerce, the direct transaction value is low in most B2B cases, and the demand seeding value is high. Someone who sees a product on a short-form video and later looks for a dealer and makes a purchase in-store is driven by social commerce, even if the attribution doesn’t pick that up.
Incrementality testing is a way to assign value to demand seeding. Use geographic holdouts with no investment and measure the downstream variance versus exposed markets. This shows the actual value of early-journey touchpoints.
Conversational commerce has a different ROI profile. The primary ROI is the savings from service costs and the increased efficiency of qualifying leads. A good system minimizes the number of unqualified inquiries coming in, therefore creating direct cost savings with a quick return. The secondary ROI is the structured intent data that becomes a data asset and enhances decision-making in other areas.
Integrated AI assistants in commerce platforms are the most easily quantifiable in terms of impact, as impact happens on a transaction by transaction basis. However, the effectiveness is contingent upon the quality of the supporting data. If data around inventory, pricing, or data regarding availability is defective, the system is likely to erode trust faster than it creates value.
How do you see AI and omnichannel commerce evolving over the next 5–10 years from a financial perspective?
In the next five to ten years, AI's main financial benefit in omnichannel commerce will be in backend, supply and distribution network, not in consumer-facing innovations. These will be necessary, but also easily replicated. Long-term advantages will come from AI in the backend supply distribution value chain. AI's direct consumer market applications have been saturated, but the distribution backend is still largely non-digitized and under automated in terms of real time data.
The firm that integrates real time supply chain data with demand, sell-through, and local search will have an unparalleled data asset. Investing in AI ceases to be an efficiency tool but transforms into a structural competitive advantage that integrates accuracy of distribution, working capital efficiency, and dealer loyalty.
The Regulatory landscape will also change the economics of AI. Compliance with data protection and AI governance frameworks will increase costs, but players with auditable and compliant frameworks will benefit. Compliance transforms from a necessary cost to a competitive advantage.
AI's consumer applications have been exhausted and backend distribution and supply value chain still needs to be automated. Competitive advantages will be long-term in supply chain networks with automation in real time data.
The established trend is what can be characterized as ambient commerce intelligence—demand, supply, and customer intent understanding in real-time, and continuously across all touchpoints. Businesses investing in building this type of architecture today, in terms of data and infrastructure commitment, will be category definers. Those postponing this will find themselves facing early movers with an irreversible data gap.
