The tech world is rife with stories of incumbents launching internal challengers to combat agile startups. One of the more telling examples is Siebel’s attempt to fend off Salesforce by rolling out CRMOnDemand.What happened? Instead of solidifying their position, Siebel created chaos. With two teams competing for the same customers, confusion reigned. One team could win, but everyone lost in clarity. Sales strategies became muddled; customers faced mixed messages, and even positioning suffered.As companies like ServiceNow look to roll out their own AI-native solutions (Flow), this history repeats itself. Two teams fighting over the same client just leads to chaos: a lesson worth heeding.Instead of launching internal competitors, incumbents might consider a hard pivot or even acquisition. Just look at what Adobe did and what Salesforce is doing now.The takeaway here? Don’t let internal rivalry cloud your strategy. Embrace clarity instead of chaos.

What's your take on ServiceNow's launch of Flow?

I’ve been using Wispr Flow for the last six months and it has become one of those products that quietly slips into your daily workflow. After a shoulder injury from playing tennis, typing for long periods became uncomfortable, so dictation wasn’t just a productivity hack for me—it genuinely made working easier.

Ironically, I stopped using it last week.

[The show, Rumble with Sinha is co-hosted in partnership with Rumble, a platform that enables one to do a podcast with their AI co-host - after all, solo talks are boring!]

Not because I found a better product, but because I ended up building my own Mac client. It does almost everything I need, and that got me thinking about something much bigger than my own setup.

I don’t think Wispr Flow has a product problem. I think it has a business model problem.

The underlying technology has become a commodity. Whether you use Whisper.cpp, Parakeet, Faster-Whisper or any of the newer open-source models, the gap between a polished commercial product and a free alternative has narrowed dramatically.

That’s not a technology problem. That’s a pricing problem.

I suspect there are still plenty of people happily paying for Wispr Flow simply because they don’t know these alternatives exist. But information arbitrage doesn’t last forever. Developers are already discovering them, building their own clients, and sharing them. Over the next year or two, I expect this to become mainstream.

If that happens, I don’t think Wispr Flow should fight open source. It should lean into it.

In fact, I’d seriously consider making the consumer product free.

That sounds counterintuitive because AI products cost money to run. But I don’t think consumers are the business anyway. I think enterprises are.

The Slack analogy comes to mind. Slack didn’t start by selling to CIOs. Individuals and small teams adopted it first because it was useful. Once enough employees were using it, companies had to buy it. Wispr Flow could follow exactly the same playbook. Let individuals get addicted to voice-first computing and let them bring that habit into their workplace. That’s where security, compliance, centralized administration and enterprise workflows become valuable enough for companies to pay.

There’s another asset that Wispr Flow has which people don’t talk about enough: voice data.

Before everyone gets upset, I’m not suggesting the company secretly sells user conversations. That would destroy trust overnight.

But what if users had a choice?

Pay for complete privacy, or use the product for free and explicitly opt in to contributing anonymized speech data for model training.

That’s a much more honest trade-off.

The reason I think this is particularly interesting is because of India. Wispr Flow has publicly said that India is one of its largest markets. At the same time, every major speech AI company wants more high-quality multilingual voice data. Today, companies are literally paying people to record speech in studios because good datasets are hard to find.

Wispr Flow already sits on top of exactly that kind of usage. Every day, thousands of people dictate naturally instead of reading scripted sentences into a microphone. With explicit consent, that becomes a valuable asset.

Maybe data licensing never becomes the biggest revenue stream. Maybe it simply offsets the cost of serving free users. Either way, I think it’s a more interesting direction than trying to convince consumers to keep paying a monthly subscription for dictation.

The enterprise biz doesn’t change. In fact, it probably becomes stronger.

No company wants its sales calls, customer emails or internal strategy documents entering public training datasets. They’ll continue paying for privacy, compliance and contractual guarantees. That’s where the money is.

Consumers and enterprises don’t need the same pricing model.

What's you're take?

Everyone is talking about Kunal Shah becoming CEO of WhatsApp.

I think they’re looking at the wrong story.

The bigger story is that Meta suddenly needs WhatsApp to work.

For years, WhatsApp has been one of the most successful products in the world.

But it has also been one of Meta’s biggest unfinished businesses.

Meta paid $19 billion for WhatsApp more than a decade ago. Yet WhatsApp still contributes a tiny fraction (less than 2%) of Meta’s revenue.

Now look at what’s happening elsewhere.

Meta is spending aggressively on AI.

Employee morale is reportedly under pressure.

The company is reorganizing teams, moving talent around, and throwing enormous resources at finding its next growth engine (Llama has failed badly).

Which makes WhatsApp more important than ever.

And nowhere is that challenge more visible than India.

India is WhatsApp’s biggest market by users.

India is one of the biggest WhatsApp Business markets.

But the monetization numbers tell a very different story.

India accounts for a massive share of WhatsApp usage while contributing a disproportionately small share of revenue.

That is the paradox Meta has failed to solve.

For years, people predicted that WhatsApp Payments would dominate Indian fintech.

It didn’t happen.

Not because the market wasn’t large enough.

Not because Indians weren’t willing to adopt digital payments.

Other fintech companies proved both assumptions wrong.

The real issue was execution and focus.

That’s why I don’t see Kunal Shah’s appointment primarily as a symbolic moment for Indian leadership.

I see it as a business decision.

If Meta wants WhatsApp to become a meaningful business, it has to crack India.

And if it can crack India, it may finally discover the monetization playbook for the rest of the world.

The question is no longer whether WhatsApp product can scale.

It already has.

The question is whether Meta can finally figure out what business WhatsApp is actually in.