AI is in a tough spot right now. While we’re seeing amazing productivity gains in many sectors, others are quickly realizing they overinvested and the returns aren’t quite what they thought they would be. It’s becoming more clear that 10X productivity gains don’t correlate with revenue. As this conflict pushes up against core business principles, executives are taking a serious look at their balance sheet, and trying to figure out exactly where they should be investing in AI.

AI in game studios generally focuses around asset creation.

For most game studios, AI is a content generator. It can mass produce art and assets (although this is probably the most hated element among players), help teams clean up their roadmaps, give community and marketing teams powerful insights, and rapidly improve development output.

And while all of these pieces bring value, they are all disconnected. Tools don’t communicate with each other. Teams build independently and rely on legacy methods to keep in touch.

A lot of this comes down to data. Analysts and project managers are the cornerstone of momentum. They keep everybody informed, help prioritize tasks, and act as the liaisons between teams.

An MCP is the glue that grounds all of your team’s efforts together.

MCP (Model Context Protocol) is the chain that connects tools and data sources to AI models. When you’re using Claude or ChatGPT, your entire chat lives in the data context of that model. That’s why when you ask Claude about your latest Steam reviews, it searches the web and browses around looking for an answer. Without clear data pipelines to source content, AI will always fall back to the general web, and in some cases artificially make data to satisfy the answer.

To solve this, you need to build a connection between your source data and the LLM you’re using. 

This is where an MCP comes in.

Without access to your data, LLMs live in a vacuum. They have no access to your Discord, Jira, Slack, game telemetry, community context… nothing.

This is where most studio execs are missing the value proposition of AI. Imagine hiring an entire marketing team, but never giving them access to anything that your company does. Of course the results would be poor.

AI is the same way. It might make your team slightly more efficient, but it’s missing 90% of its power without clear access to good data.

The traditional data pipelines necessitate talented analysts, lots of code, and too much time. It’s blocking your entire team.

Making great product decisions requires a special combination of time, data, and intuition. It’s a three-legged stool, and it won’t work if any one leg is missing. Most teams operate with a traditional data knowledge gap. They are sitting on mountains of data points, but being able to understand it is where they fall short. Fortunately, AI has an incredible capacity to ingest and understand data at unfathomable scale.

‍

Each team has unique data chokepoints that are not allowing them to work at the pace they could.

Production
Too many tickets, too many priorities coming from countless directions. If you want to build a product or community-driven team, PMs are the decision makers that keep the ship moving in the same direction. And in the age of AI, you can build in the wrong direction just as fast as you can the correct one. But without clear knowledge and understanding, getting on the right track is very difficult.

Marketing
Marketing and user acquisition teams are already loaded with data. The problem is that it’s split across too many platforms. Between social listening, paid acquisition methods, and organic content pushes, it’s nearly impossible to find connections between them all. So they fall back to reading things manually and crossing their fingers that they are making the right call.

Community
Community managers are the frontline in games and community-driven products. Their task is also dramatically underestimated. How hard can it be to keep up with Reddit and Discord? Well, when you are getting 500 messages an hour, there’s literally no way a human can possibly keep up, let alone be able to accurately deliver reports to the team to make critical calls.

Live Ops
Fortunately, there are a lot of great technical solutions in the live ops space. Where things currently fall short is the marrying of community sentiment with live ops alerts. Catching exploits, bugs, and cheaters before they go viral can be critical to a game’s long-term success.

Leadership & Strategy
Even the most talented teams will flounder in a directionless sea. Making those critical decisions can be make-or-break for a studio’s success. And while there are many talented and visionary leaders out there, the best ones know how to push back on their intuitions and marry their ideas with tangible results and data. And at the same time, the last thing you want to do is pull in your most powerful talent from their important tasks just to help you with a research project. Access to good data at an executive level ironically often goes ignored, even though it can be the most consequential.

‍

An MCP unblocks every team at once.

Between time, data, and intuition, data is easily the most solvable. The unlock here is that data needs to be accessible. This is where an MCP comes in and saves the day, or in this case, the entire product!

Let’s look at a simple, yet valuable, example. 

You’re a producer and you’ve been tasked with figuring out how we can reduce churn among new users. There are many ways to approach this problem. You can talk to UX, ask your data team to see what they know, maybe talk to devops and see if there are technical issues that are occurring. All of this is going to take dozens of hours of discovery, research, and effort from already pressing tasks.

With an MCP, you can retrieve data-driven results in minutes. 

Because your data has been consolidated between teams, and is accessible via natural language, you can start with a simple prompt. 

“I need to figure out how bad our churn problem is for new users. Can you help me measure it, and look at various customer-facing endpoints to find common patterns or problems they may be encountering?”

From there you can refine the results. Maybe look at messages from users less than 30 days old. Or explore email threads from new customers to see where they are having the most difficulties. 

And in less than an hour’s worth of work, you can have a concrete measure of truly how bad the problem is, five areas where it can be most improved, and what each team member can do to help.

It’s this kind of access to data that truly unlocks teams to move faster. You can make clearer, stronger decisions, and build products that align with your business goals and your customers’ needs.

How to unlock the capability of AI across your studio.

Fortunately, getting an MCP set up for your organization isn’t too difficult. There are a small handful of organizations that offer it, just like you would see in software as a service. They all generally offer the same service: solid connections to various tools, levels of support and service, and sometimes other nice features that can make indexing enormous amounts of data a bit faster. The largest differentiating factor is figuring out what tools you need data from. 

Industry-specific tools will be more likely to support the necessary platforms. In games, for example, you’ll likely need solid connections to Steam and Discord. Games-specific tools, such as Kinn, will also be familiar with the workflow and what formatting the data needs to be in for each member of the team.

Setup and maintenance are incredibly easy, and can often complete in a single day. From there, you can add additional tools you might need, or configure role access and use. It can be important to monitor costs early on to get an understanding of token use and volume over time.

With Kinn, studios can have an MCP running across multiple sources in less than an hour.

Kinn, at its most basic level, is a secure endpoint for the tools that game studios and community teams need. We have secure APIs, MCPs, and connectors to dozens of tools such as Discord, Steam, GitHub, Slack, and Reddit. 

Since most content is public, connections can be established in minutes, and data will start flowing immediately. For locked content such as Discord or Slack, simple bot installs take a few clicks and administrative permission.

From there, you can invite anybody on your team, and they are each equipped with pre-built queries and solutions for their specific role.

Our expert team works with a variety of roles in studios and coordinates with each of them to learn their existing workflows, where their existing data chokepoints currently exist, and unblock those as quickly as possible. 

Over time your entire team will be more knowledgeable, more agile, and shipping faster than ever.

Explore our playground environment to try it for yourself, or schedule a demo with our team to see how Kinn can help your team today.