While the appeal of chatbots and conversational commerce is undeniable, I’m highly skeptical about their future in bot stores built on existing messaging platforms such as Messenger, Kik, Telegram, and Slack, even after taking into account the massive infrastructure, reach, and distribution these offer developers.

Since the bot craze started last year, the industry has created thousands of bots  —  mostly meaningless and glorified versions of IVR systems —  across an assortment of messaging platforms. Whether we’re booking flights or checking the weather, we have bots for almost every use case. Even POTUS has a bot now!

As per recent estimates, there are at least 11,000 bots on Messenger and 20,000 bots on Kik.

You can understand why all the existing platforms are pushing aggressively for bots as they want to win the race and become the next app store equivalent. However, are these platforms really sold on the potential of bots? Do they really understand bots’ strengths and weaknesses? Have they put real thought behind what a bot store should look like? Are they providing the right support and tools to the developer community?

It’s more likely that their efforts are driven by a desire to showcase themselves in the eyes of the market and the developer community. It’s no surprise that Telegram, Kik, and Messenger came out with their bot stores in close succcession — so as not to allow anyone to become the de facto leader. The same motivation is probably behind the recent launch of similar features from Facebook (quick replies) and Slack (buttons).

I’m not saying things won’t improve in the bot ecosystem or on messaging platforms, but it’s good to take a step back and be aware of the reasons developers might burn their fingers if they approach bot stores as they would app stores.

First of all, discovery of bots on any messaging platform is very limited right now. There’s hardly any organic way for developers to get in front of massive audiences in bot stores. This was the single most important pain point raised by the bot developers at a recent event. The best reply I heard from a platform lead was “So far, we’re relying on developers themselves to leverage their existing audience.” But this attitude defeats the single most important motivation to develop bots on existing platforms.

Bots offer great promise, but the very problem they intend to solve (e.g., a universal conversation flow) comes into question if developers are banking on individual messaging platforms. For example, to get a bot to understand user preferences, developers would need access to personal data from Facebook, Slack, Twitter, LinkedIn etc., yet it’s very unlikely that each of these platforms would to make their data available to a competitor. For reference, look at how much compatibility iOS and Android offer and imagine that for five to six leading messaging platforms. It could get dirty soon!

Even if data was easily accessible across platforms, there’s a great degree of effort required in maintaining bots across diverse platforms. Think about the effort it takes to maintain apps across two leading app platforms, Apple and Google. With multiple messaging platforms having their own bot kits and stores, approval processes and stringent requirements  — even with help from several bot deployment platforms  —  cross-platform bot management will soon become a huge pain.

In terms of control — unlike app stores where developers have access to native systems to define their app’s experience — in bot stores, developers are left waiting for the messaging platforms to define the interface and move desperately needed features through the pipeline. This makes it impossible for bot developers to control the end to end experience of their users, who will soon start churning out.

Moreover, Apple and Google provided robust payment infrastructure to app developers, allowing them to create solid paid, freemium, or ad-supported business models. In the case of bot stores, no one currently provides any kind of sophisticated payment system that would allow bot developers to start thinking about not just the novelty of bots but how to shape them into a solid revenue generating business.

Developing and deploying a bot is relatively easy compared to building a native app on iOS or Android. This will soon lead to a proliferation of bots and a crowded space for developers  —  more congestion than we face with apps right now. In order to stand out, developers will start falling into the trap of distribution via paid in-bot ads or featured listings in the stores, spending more and more money to reach out to their target audience.

I believe that bot developers can save themselves from pain down the road by using existing platforms as test beds for their initial bot experiences but then carefully evaluating if it makes sense for them to break free and develop their unique bots and businesses independent of the big platforms.

This post originally appeared on Medium.