In all the excitement and investment in the latest flavors of AI – namely generative and conversational AI – one looming question is revenue models for products like the recent Microsoft Copilot As we’ve speculated, there will be a combination of enterprise (B2B) software licensing for consumer-facing products (B2B2C) and freemium apps (B2C).
OpenAI for example is doing some combination of all the above, and that will continue to develop – especially as it rolls out its app store. Another influential player that could set the tone is Microsoft. Its Copilot, formerly known as Bing Chat, continues to roll out, including physical keyboard buttons.
After speculating its revenue model recently on our podcast, This Week in Local, the latest puzzle piece was placed this week for CoPilot’s revenue model. Though we predicted it would be more of a loss leader to entice users into Microsoft products (e.g., Office 365), Microsoft has a more dimensional strategy.
Specifically, Copilot will offer a Pro plan. Similar to OpenAI’s app, this will include paid upsell that unlock features (the freemium approach noted above). At $20 per month, it includes expanded generative AI features on PCs and Macs across products like Word, Excel, PowerPoint, Outlook, and OneNote.
E52 Explores MSFT’s New Co-Pilot and 2023’s Grim Toll on Startups
Boost & Broaden
Fleshing out CoPilot’s approach a bit further, it’s not just a freemium app (again, OpenAI’s approach with ChatGPT), but an integrated capability. In other words, it’s a paid feature that cuts across several products. This not only boosts and broadens its value proposition but gives it a bigger shell to grow into.
For example, use cases Microsoft lists at launch include summarizing text in Word, organizing data in Excel, and drafting emails in Outlook – all through natural language prompts. Users also get 100 daily “boosts” – basically, credits – for generative images in Designer (formerly Bing Image Creator).
Speaking of generative art features, Copilot Pro will provide speedier work, improved image quality, and image editing options. The Pro upsell also boasts early access to the latest generative AI models, such as OpenAI’s GPT-4 Turbo, and server priority for better performance during peak hours.
As noted, all the above will cost $20 per user per month, which is signifigant in its comparison to Microsoft’s core offerings like Microsoft 365. For example, it’s a meaningful price boost from the $6.99 Microsoft currently charges for 365 Personal Plus. Having them together brings the total to $27.
Given the weight of that upsell (almost a 300% larger price tag), it will likely resonate most in the near term with professionals and enterprises (I suppose that makes its branding on target). It’s available this week for companies subscribed to 365 Business Premium, 365 Business Standard, and a few other tiers.
Table Stakes
Altogether, Copilot could serve Microsoft as both a means to boost the value proposition of its software suite, as well as a revenue center in its own right… and a robust one given the pricing. This puts Microsoft in a strong position in AI, and that can be seen in Wall Street’s reaction among other places.
In that sense, it makes sense that Microsoft (and OpenAI for that matter) are cultivating B2B and B2C freemium models. Though we speculated that Microsoft will treat CoPilot as a loss leader to boost appeal for Microsoft products – and we stand by that – there’s also proven end-user demand to pay for AI.
Microsoft knows this and sees an opportunity to tap into heavy demand signals for AI within its core realm of business productivity. In fact, it has reported that 40 percent of the Fortune 100 participated in the Copilot early access program. We’re talking capabilities that could become enterprise table stakes.
Making matters better for Microsoft, there’s a strong product-market fit. Automation fits well into its productivity suite – everything from entering the right Excel formula to building a pitch deck (insert Clippy joke). Here it has chosen the right branding with “Copilot,” as we’ve discussed on TWiL.
Beyond that, the Microsofts and OpenAIs of the world have to feed the beast. As you probably know, the server loads in running large language models and generative AI in the cloud can get pricey. So to realize worthwhile margins – beyond those inherent in SaaS – they’ll have to attack from many angles.
