Should you use AI to help plan or pitch a game project?

Recently, when clients came to us for game development or game marketing proposals, some asked whether proposals and plans made with AI are now complete enough. Can a proposal like this really get the project done and lower the overall cost? Put simply, can you just follow what the AI produces and get a solid result?

An AI proposal is one option, but there is a lot to watch out for.

Proposals are faster now, but so are the unknowns in actually delivering them

From years of client work, GameNobility knows one thing well: when clients evaluate vendors for a project, they don't always talk to just one company. Comparing two or three proposals on price, content and value for money is normal. That is completely reasonable, and we welcome comparison.

Since AI came into wide use, one problem has been showing up more and more.

AI-generated proposals tend to read as very complete. Full feature lists, clear logic, and at first glance almost no weaknesses. The problem is that this content is generated from "the vendor PM's description of the requirements". It describes an ideal situation, not your actual budget, your team's resources, or the goals your project really needs to reach.

The result is that every feature in the proposal looks doable, but the real cost of building it may already be over budget, or it may not be feasible at all. Nobody points this out at the proposal stage.

How to spot an AI-generated proposal

Before you decide, you can review the proposals in hand from the following angles.

Content and structure

The content of the proposal itself is often the most direct clue:

Are there relevant case studies to back it up?
Does the whole structure look too easy to deliver (that is, the vendor raises no real-world concerns)? Experienced vendors usually point out a few areas of concern, such as how users will log in so you can collect valid data.
Has the user experience been considered, such as the overall flow, likely real-world experience issues, and even the devices and situations users will play in?

Technical feasibility

Some technical promises seem reasonable at first glance but don't hold up:

Claims to support many platforms (PC, iOS, Android, even Switch and PlayStation), but at an unusually low price: there is no explanation of the different porting costs for each platform. In the industry, that pricing structure is impossible.
The art style is described as "custom design", but there are no concrete style references, development process or matching prices: custom art needs full planning based on the client's brand story. It can't be covered in one sentence.

Understanding of the client

A good proposal should show that "they really understand your project":

The proposal contains none of the keywords or background the client mentioned: it feels like you could swap the logo and use it for any client. It lacks focus, such as analysis of the goals the client wants to reach or what this approach can achieve, even though every design choice should have a clear purpose.
It never says "what we won't do": anyone who truly understands a project will clearly state which features are outside this phase. A proposal that says it can do everything is the one to watch out for.

A real example

We once saw an optimization proposal for a branded shopping game. It opened with: "Keep the core structure, streamline development, and ensure future scalability."

It sounds professional. But read closely and every work item has exactly the same format, with no sense of priority. The leaderboard system says "recommend postponing or simplifying", the mini games say "not developed in this phase, space reserved", and the conclusion says it "can reduce overall development costs by about 20% to 40%", with no calculation behind it at all.

More importantly, swap out the client's brand name and this proposal could be used almost as-is for any client who wants a branded game. It doesn't answer the client's real questions. It just wraps a generic framework as a "custom solution". To most clients, custom means a full plan built around the needs they raised, not a generic framework.

Another situation is a proposal that looks very complete, with clear work items, distinct phases and a timeline, but when you reach the end you find it has no pricing at all. The more detailed the features, the harder it is for the client to judge whether the plan fits their budget, and the harder it is to compare proposals in a meaningful way.

In both cases, the client only finds out something is wrong during development.

With a limited budget, can you really make five games?

One client wanted to develop five mini games with a budget of NT$500,000. An AI-generated proposal might say that number is "feasible", but in reality, the UI design, programming and QA testing for even one game take considerable labor. Five games within that budget is almost impossible to deliver. It's not that they can't be made; it's a question of what the result will look like.

A quote below a reasonable market rate usually means mass-produced AI visuals, code built on an off-the-shelf framework, and a game experience with almost no tuning. Forgotten after two minutes of play, no connection to the brand, KPIs not met: the industry calls this a disposable game.

When something is made just to use up the budget, what gets used up is not only the budget but also what the brand could have built through the campaign.

When a proposal's price is far below a reasonable range, the question to ask is not "I found a cheap one" but: is what this price buys really what you need?

Behind the budget question is really a decision question.

How to turn gamification from a marketing expense into a brand asset worth serious investment deserves its own discussion.

AI art: it looks doable, but it can't actually be delivered

One common but unreasonable promise in the market today is: "All the art will be done with AI."

It sounds cheap and efficient, but AI still can't keep a character's style consistent and unified. In game development, the same character has to appear in different scenes, expressions and poses. That needs human artists to set up full character design guidelines and check style consistency frame by frame. Using AI art for the whole project means using a tool that can't keep a memory to do professional work that needs a high level of consistency.

But the more fundamental problem is planted back at the quoting stage.

A reasonable art proposal must get both sides to agree on the following before the project starts:

Scale: how many scenes, characters and objects are needed
IP design: whether original design is needed, or an existing IP will be used
Asset integration: whether the client's existing design assets need to be used and integrated
Revisions: the maximum number of revisions for each work item

Only from these conditions can a truly custom quote be made based on the budget and industry rates, with work items added or removed as needed.

This process needs an experienced person to sit down with the client and confirm each item. If it isn't spelled out before the project starts, the two sides often have very different ideas of "how far the art will go". When the problem shows up mid-project, the outcome is either more budget or lower quality, and neither is what the client wants.

Being clear about what can and can't be done is real professionalism

Many clients reviewing a proposal don't fully understand the technical meaning behind each feature.

For example, a client wants a leaderboard and player data records in the game. These sound like ordinary game features, but they actually involve building a full back-end system and admin panel, with engineering costs that have a market rate. If a proposal just lists "leaderboard support" without explaining how much engineering it means or how it affects the budget, the client has no way to judge whether the proposal is reasonable.

When we talk with clients, we try to explain in a way people outside the industry can understand: what this price can achieve, and which features won't fit in this budget.

When the budget is limited, we don't just say "we can't do it" and stop there. Instead, we:

Offer a simplified option and explain the real difference between the simplified and full versions
If an essential feature can't be cut, analyze which game features or presentation are secondary and reorder the priorities
Let the client decide with a full understanding, instead of finding out after signing that features they thought were included aren't there

"We won't build a leaderboard" or "We can do everything": which is more believable?

Honestly, when we assess a budget and clearly say "a leaderboard can't be included in this budget", the client may receive another proposal at the same time claiming the same budget can cover a leaderboard, player data records and multi-platform support, with a long feature list and an even lower price.

At that point, the question is not "which one is cheaper" but: can this proposal actually be delivered?

Features like leaderboards and player data records need a full back-end system, with engineering costs that have a clear market rate. If a proposal includes all of them but the price is far below a reasonable range, there are only a few possibilities:

These features are only "written in the proposal" and will shrink or disappear during development for one reason or another
The technical quality is very poor, and problems keep coming after launch
Money or staff run out halfway through, and the project simply stops

We say clearly "no leaderboard within this budget" because we understand the cost of every piece of engineering and take responsibility for what we promise.

Whether something can be done and whether someone is willing to write it into a proposal are two completely different things.

AI doesn't know who the client is, but an experienced team does

The deeper issue is how well the vendor knows the client.

AI generates content from a description of the requirements, but it doesn't understand you as a client. Two companies may both want a game, but one wants brand exposure, another wants to drive traffic to a campaign, and another is aiming straight at revenue. Different starting points lead to completely different development directions, feature priorities and budget allocations.

AI won't ask about these on its own. It doesn't know the client's industry, the business logic behind your project, or whether anyone on your team can support the work afterward. It will only present the requirements you gave it as completely as it can.

A proposal that is logically complete may still be heading in the wrong direction for the client from the very start.

In all our years of client work, we always talk properly with the client before doing any planning. Not just listening to requirements, but understanding: What is the client thinking? Why do they want a game at this point? What problem is this game meant to solve for them?

Only when these questions are answered will the proposal be truly tailored to the client, rather than a document produced from a template.

We're not afraid of being compared, but we worry you'll choose wrong

Honestly, we're not afraid of being compared with other proposals.

What really worries us is something else: a client sees a complete-looking proposal with an attractive price, decides without further checking, and only finds out during development that the cost of those features was badly underestimated, that the timeline was never realistic, or that the whole direction never matched the client's real needs.

We have seen this happen many times. Some of these projects came to us to take over, but many could no longer be continued or had to be torn down and rebuilt.

A lack of care at the start led the client to the wrong decision. When the project can't be delivered, what's lost is not only money but also time, and confidence in using games at all.

A proposal is only the start; what matters is whether it can be built

It's true that AI makes proposals faster and better-looking. We use AI ourselves to help organize documents and speed up our work. But AI is a tool, and tools need people to check and review their output.

What really has value is not a proposal that looks complete, but someone who can tell you clearly at the proposal stage how much development a feature really takes, how far this budget can go, and whether this direction really suits your industry.

That judgment can only come from people who have done enough projects and truly understand game development.

If you are evaluating a proposal right now, we have just one suggestion: before you decide, ask one more question. Can this proposal really be built, or does it only look like it can?

Or bring in a company with years of experience to compare proposals, so both sides can talk face to face and clarify the details behind the proposal.

Have any questions about game development? Come talk with GameNobility

FAQ

Q1: I already have a proposal from another company. Can you take a look at it for me?

Yes. This is one of the requests we get most often. You don't need to pay or commit to anything first. Just tell us what the proposal says and your budget range, and we'll give you a straight opinion, including what is reasonable and what needs to be clarified.

Q2: I know nothing about game development. Can I still come to you?

Of course, and you are exactly who we most want to help. We are used to talking with clients in plain language, so you won't sit there listening to jargon you don't understand. Just tell us clearly what you want to achieve, and we'll tell you the most sensible way to do it.

Q3: My budget isn't big. Will you take the project?

The budget itself isn't the issue. The issue is whether the budget matches the requirements. We won't hand you a careless proposal because the budget is low, and we won't agree to things we can't deliver just to win the job. If, after our assessment, your budget really can't support what you want, we'll tell you directly and explain what can actually be done within it.

Q4: Another company started developing for us and ran into problems halfway. Can you take over?

It depends, but we have taken over quite a few projects like that. Before taking over, we first assess the existing code and architecture, confirm whether work can continue and what it will cost, then explain it clearly to you. Not every project can be saved, but we'll give you an honest assessment and won't exaggerate just to win the job.

Q5: Are AI-generated proposals completely useless?

Not completely. It depends on whether someone is checking the work behind them. AI can quickly organize a structure and generate documents, but it can't judge for you whether the budget is reasonable, whether the features can really be built, or whether the direction fits your business goals. In a good proposal, AI can be a tool, but it can't be the only author.

Q6: The game I want is very simple. Does it really need all this process?

A "simple game" in a client's mind and in a developer's hands are often two different things. Many clients tell us they want a "small" game, but as we talk we find it includes a login system, a leaderboard, member data integration... and each of these is an engineering cost. We're not trying to scare you. We just want to make these clear before starting, so there are no surprises later.

Q7: What kinds of game projects have you worked on?

We have worked on branded interactive games, marketing campaign games, web mini games, simulation and nurturing games, shopping game integrations and more. The development logic and cost structure differ greatly between game types, which is why we assess every project from scratch instead of using a fixed price list.
If you have an early idea for a direction but aren't sure how to set the specs and budget yet, talk with us about your needs first. After one round of discussion, the direction usually becomes much clearer.

Q8: I only need programming or art. Do you offer individual services?

Yes. Besides full game development projects, we also offer individual professional services. You can commission programming only, art only, brand visual design, game script planning and more on their own.
This suits clients who already have some resources and only need to fill a specific gap. Whether you already have the game design or art and need someone to build the programming, or the other way around, we can work with you either way.