From device leasing to ₹9 crore: the vertical AI pivot

SwishX raised money, built a device-leasing business for pharma, and then shut it down. The pivot wasn't desperation — it was diagnosis. The team saw pharma sales reps drowning in manual workflows, recognised they already had the customer relationships to fix it, and rebuilt as an agentic AI platform targeting the pharma CRM software market that incumbents like Veeva Systems and Salesforce have largely ignored at the mid-market and India tier. They've since hit roughly ₹9 crore ARR at seed stage and are targeting ₹42 crore by FY27.

Why the Device-leasing Model Had to Die

Capex-heavy models punish founders twice: once when you buy the inventory, again when a customer churns and you're left holding hardware. Swish Club — SwishX's previous identity — leased devices to pharma field reps. Every new customer meant more upfront spend. Margins were thin, differentiation was near-zero (any competitor with capital could replicate it), and the model scaled linearly with headcount and hardware, not with software leverage.

The founders didn't need a consultant to tell them this was broken. The signal was already inside their customer base: pharma sales managers were manually reconciling distributor orders in spreadsheets, field reps were chasing approvals over WhatsApp, and no one had clean visibility into what stock sat at which stockist. The problem was screaming. The device-leasing revenue was just noise.

This is the pattern that separates a real pivot from a rebrand: you don't invent a new customer — you solve the adjacent, painful problem the customer you already have is complaining about.

The Wedge-pivot: Follow Your Customer's Pain, Not the Trend

The wedge-pivot playbook has three moves. First, kill the undifferentiated business before it bleeds you dry — not after. Second, map the specific, painful workflow your existing customers are still doing manually. Third, own that workflow so completely that switching to a generic horizontal tool feels like a downgrade.

SwishX's wedge is pharma's commercial back office: secondary sales tracking (what moves from distributor to chemist), field force automation (daily call reports, tour plans, sample management), and distribution visibility. These aren't glamorous problems. They're also not solved by asking a generic AI copilot to summarise a Salesforce dashboard.

Manesh Naidu, Chief Commercial Officer at Tris Pharma, put the principle precisely:

The wedge-pivot only works when the problem is specific enough that a generic tool can't fake it. Pharma's distribution chain — manufacturer to C&F agent to stockist to chemist — has compliance rules, shelf-life constraints, and trade scheme structures that a horizontal copilot has never been trained on. That specificity is the moat, not the AI itself. This is also why building a genuinely differentiated AI product looks nothing like the AI-theatre that fooled investors for years — the wedge is in the domain knowledge, not the model.

Vertical AI Beats Horizontal Copilots — Here's Why

The global life science software market is projected to reach roughly ₹1,88,000 crore in 2026 and grow to ₹3,20,000 crore by 2031, according to Mordor Intelligence. The incumbents — Veeva Systems with 26.81% market share, Salesforce with 16.4%, Oracle with 11.21% — are all moving upmarket and toward enterprise. Veeva just signed 125+ customers onto Vault CRM and accelerated its legacy CRM end-of-support. Salesforce signed AstraZeneca and Novartis within two months of launching Life Sciences Cloud.

None of that activity is aimed at a 200-rep Indian pharma company running its secondary sales on Excel.

Vertical AI wins in the mid-market for a structural reason: the domain-specific training data, compliance rules, and workflow integrations required to serve pharma's commercial operations are a meaningful barrier. A generic horizontal tool can answer questions about your CRM data. A vertical platform that already knows how Indian pharma distribution works — C&F agent hierarchies, chemist beat plans, stockist credit cycles — can automate the decisions, not just summarise them.

ZS Associates understood this early. Their ZAIDYN Commercial platform embeds life-sciences-trained agentic AI into field planning, customer engagement, and incentive management. IQVIA launched IQVIA.ai in March 2026 with 150+ deployed agents and over 100 AI patents. The race isn't between AI and no-AI — it's between platforms that know pharma's commercial workflows and those that don't.

The Market SwishX is Entering is Large and Contested

The pharma and biotech CRM software market is growing at a significant pace, per Research and Markets. The India pharma contract sales organisations market alone is projected to reach ₹5,000 crore by 2030 at 12.5% CAGR, per Grand View Research — and that's only the outsourced sales layer, not the broader commercial software stack.

SwishX's ₹9 crore ARR is a rounding error against those numbers. That's not a criticism — it's the point. The wedge-pivot strategy deliberately starts small and specific, then expands. The danger isn't the market size; it's losing the specificity that made the wedge work in the first place as you scale.

Avinob Roy, Vice President and General Manager of Global Commercial Product Offerings at IQVIA, flagged the scaling challenge directly:

For SwishX, this means the path from ₹9 crore to ₹42 crore ARR isn't just about adding customers — it's about maintaining the domain depth that makes the platform defensible as the team grows.

Distribution Beats Over-engineering — but Only with Specific Positioning

Here's a trap founders fall into after a successful wedge-pivot: they over-engineer the product instead of scaling distribution. The product that got you to ₹9 crore ARR is probably good enough to get you to ₹42 crore — what changes is how many of the right customers know it exists.

But distribution only compounds when the positioning is painfully specific. A message like "AI for pharma" is noise. "Agentic AI that automates your secondary sales reconciliation and field force daily reports" is a message a pharma sales director can forward to their CEO. The ICP (ideal customer profile — the exact type of company most likely to buy) has to be locked before you scale the sales motion, or you just amplify confusion.

This is the nuance the wedge-pivot playbook demands: distribution speed and product specificity have to move together. Push distribution before the positioning is sharp and you get demos that don't convert. Sharpen the positioning first and distribution compounds fast — every happy customer becomes a reference in a tight industry where everyone knows everyone.

This dynamic is also why most SaaS funnels leak at the positioning layer, not the traffic layer — you can drive all the right visitors to your site and still lose them if the message doesn't match the specific pain they walked in with.

The Boring-business Truth About Cloning and Undercutting

SwishX's pivot is a version of a pattern that practitioners who've been in the grind long enough recognise: find a proven, painful workflow, build to feature parity with whatever manual or legacy process the customer is using today, and price it so the ROI calculation is obvious. It's not glamorous. It's effective.

The most durable vertical SaaS businesses aren't built on novel technology — they're built on the willingness to go deep into an unglamorous workflow that a horizontal platform will never prioritise. Pharma's secondary sales reconciliation is exactly this kind of problem: painful, repetitive, compliance-sensitive, and completely ignored by the enterprise vendors chasing AstraZeneca.

The philosophical tension here is real. Chasing a ₹42 crore ARR target by automating chemist beat plans is not the same as building the next Zerodha. But the founders who've shipped multiple businesses — and paid the tuition on the ones that failed — tend to converge on the same conclusion: boring problems with recurring pain produce recurring revenue. The ₹9 crore ARR SwishX has already built is more real than the pitch deck of a hundred more ambitious ideas.

This is also why the DoableClaw blog-engine architecture review lesson applies here: once you've diagnosed the real problem (in SwishX's case, the capex model was broken and the workflow pain was the real opportunity), more analysis is just a comfortable way to avoid the build. The founders who succeed are the ones who stop reviewing and start shipping.

API Dependency Fears are Mostly a Distraction

A recurring objection to any AI-native startup is the dependency risk: what happens when OpenAI changes its pricing, or a competitor's LLM gets better? For a vertical AI platform, this concern is largely misplaced.

The defensibility of a vertical platform like SwishX is not in the underlying model — it's in the pharma-specific workflow logic, the compliance rules encoded into the agents, the integrations with distributor ERPs and field force tools, and the training data that accumulates as customers use the platform. Switching the underlying LLM is a backend decision. Rebuilding the domain knowledge is a years-long investment.

Practitioners building real AI products understand this: you architect for LLM flexibility from day one, so swapping providers is a configuration change, not a crisis. The real lock-in is the workflow, the data, and the trust the field team has built with the recommendations the platform produces.

What Sapre Actually Said — and What It Means for Your Business

Dushyant Sapre, Founder and CEO of SwishX, has made a bold claim about AI's potential to streamline pharma commercial processes. That claim is aggressive by any measure, and it's worth examining what it actually implies rather than just repeating it.

The processes most at risk aren't strategic ones — they're the manual reconciliation, the report generation, the approval routing, the data entry that currently consumes a pharma sales manager's week. If agentic AI handles a significant portion of that administrative layer, the sales manager's job doesn't disappear — it shifts toward the work that requires judgment, relationships, and market intuition.

Derek Choy, Head of Product at PharmaForceIQ, frames the governance requirement that makes this work in practice:

For a pharma sales director evaluating SwishX, the real question isn't whether AI can eliminate manual processes — it's whether the platform has encoded your compliance rules, your brand constraints, and your approval hierarchies before the agents start acting. A platform that automates without governance creates a different kind of mess.

The Governance Layer Incumbents Keep Missing

The reason field teams resist AI recommendations isn't scepticism about AI — it's that the recommendations arrive without context. A rep who gets a next-best-action suggestion from a black box has no reason to trust it over their own judgment built from years in the territory.

The vertical AI platforms that win field adoption will be the ones that make every agent action traceable: why did the system recommend this doctor, this channel, this timing? That transparency is not a nice-to-have — it's the adoption lever. A pharma rep who understands why the platform is suggesting a particular call sequence will use it. One who doesn't will ignore it and keep working from their own call list.

This is where SwishX's vertical focus gives it a structural advantage over a horizontal copilot: the governance layer can be pharma-specific from day one, with compliance rules and brand guidelines encoded before the agents act, not bolted on after a compliance team raises a flag.

What Founders Should Do Before Their Next Pivot

The SwishX story is a clean case study in wedge-pivot discipline, but the lesson only applies if you do the diagnostic work first. Before you kill your current model and rebuild, three questions need honest answers.

First: is the problem you're pivoting toward already validated by the customers you have — or are you projecting? SwishX had pharma customers who were already complaining about the workflow problem. That's a demand signal. A pivot toward a problem you've only read about in a market report is a different risk entirely.

Second: is the new problem specific enough that a horizontal tool can't fake it? If your answer is "we'll add AI to [generic workflow]", you're building a feature, not a platform. The wedge has to be narrow enough that domain expertise is genuinely required.

Third: can you get to meaningful ARR before the runway runs out? SwishX raised ₹19 crore in seed funding and has ₹9 crore ARR to show for the pivot. That's a real validation signal, not just a rebrand.

Conclusion

The wedge-pivot only works if you follow a real demand signal — not a market report, not an AI trend, not a rebrand. SwishX had pharma customers already complaining about the workflow problem before it rebuilt. If you're sitting on a capex-heavy or undifferentiated model right now, the question isn't whether to pivot — it's whether the specific pain your existing customers are describing is narrow enough to own completely.