AI and CAD

Let’s start with the idea of a voice-activated, artificial intelligence (AI) CAD tool.

Imagine a group of people in a room planning out a new product, a pair of garden shears. There is an industrial designer named Dieter Rams, an engineer named Nikola Tesla, and a marketer named Mark Eter.

The meeting starts with Mark, laying out the goals and information based on market research:

Mark Eter

“The price point is $25, it has to be able to withstand 100 lb of grip pressure and a drop from 12′ high. They will be sold to Home Depot, and will have to sit along those 4 pictured garden shears I emailed yesterday. Let’s hear some ideas”


“I have a concept in mind.”

Hearing this, the CAD tool creates a new folder, and starts a file named “Concept 1”.

Dieter starts sketching out a concept on his touch screen laptop in his CAD tool. While sketching, Dieter proclaims,

Dieter Rams

“Good design must make a tool useful, so these shears need to help someone be better at trimming their plants. I’m picturing a measuring tape that pulls out of one of the handles, so that people who want to keep their plants trimmed at specific lengths can quickly measure, then cut. This will require making the bottom handle much bigger and bulkier than the top, but I think it will still look good.”

As he finishes his comment, he has free formed sketched the basic parts and shares it with everybody.

Mechanical Engineer

The engineer, Nikola, shares his thoughts,

“I don’t think this concept will be easy to manufacture. While the design makes the bottom bulkier, it still doesn’t provide enough room for a round spool of measuring tape due to the wall thickness that will be necessary. Let’s start a second concept by starting with a copy of this.”

The AI software picks up his words, makes a file called “Concept #2” and includes the same sketch.

Mechanical Engineer

Nikola points to the bottom leg, and the sketch and says,

“Can we erase that portion on bottom, and redraw a more circular bump?”

Dieter does so in response. The design now works more functionally, but doesn’t look great any more.

Mark Eter

“I’m concerned.”

Mechanical Engineer

“It will fit our price point.”

Dieter Rams

“But it’s ugly. No one will want it.”

Mechanical Engineer

“Perhaps we’re thinking in shapes that lend themselves better to injection molding, let’s think about handle shapes that can be extruded.”

Mark Eter

“That may be great from an aesthetic view too, because the competitors all have a similar look due to being molded. This will differentiate us.”

Dieter Rams

“Alright, let’s start concept #3.”

Hearing this, the software, creates a new file named “Concept #3”

Dieter Rams

“Let’s start with the main leg sketches from Concept #1”

The software copies these from Concept #1 to Concept #3.

Dieter Rams

“Now let’s extrude them to a depth of 1/2”, says Dieter.

The software picks up on that and extrudes them to the proper depth. The design is taking shape.

Mechanical Engineer

“What if we had holes or squares removed from those extruded handles for weight and material savings?”

The software creates a hole pattern in the handle as a starting point and creates holes in concept #3 and square cuts in a new concept #4.

Mark Eter

“How do we think we’re doing with all the concepts? Can we look at them all at once?”

The software responds by throwing up a split screen of the 4 concepts that are being worked on.

Let’s break from our example for the moment.┬áThe above scenario is only some of the capability that would be possibly by combining AI with CAD.

For the last 6 years, I’ve heard how cloud computing is the future of CAD from the software producers. Based on hundreds of customer conversations I’ve had, 250 last year alone, the customer base isn’t convinced if all they get is the same functionality, with less administration.

Cloud based software has become a focus as many companies prefer the SaaS (Software as a Service) model due to the steady revenues and data control by the provider (harder to switch). What about the value to the customer?

  • automatic updates
  • disaster recovery
  • lower up-front capital costs
  • bandwidth flexibility
  • increased collaboration
  • increased security

Those points are debatable for any organization. If shifting an existing software to a cloud version new possibilities the cloud provides must be consider, such as AI, otherwise adoption will be less than stellar. Here are features the cloud can provide that regular software can’t:

  1. Insight from shared data
  2. Increased functionality that isn’t possible on basic hardware

Since people have different voices, or choose different words for commands, shared data collection allows for the software to learn and adapt using AI algorithms, along with some human interpretation. The net benefit is that the software continues to improve and adapt for customers as a result of the shared information that wouldn’t be possible on the proper scale without the cloud. Also, to pull off this type of software, quick, powerful processing must happen, and that is enabled through infrastructure and equipment that wouldn’t be worth the cost of replicating at each customer facility making it ideal for the cloud.

It’s time to provide something new, and that something is CAD powered with AI.

Value Proposition

for End User

  • Software behaves as an additional worker and increases efficiency
  • Keeps files managed automatically
  • Allows direct interaction from the customer without knowledge of software
  • Allows the user to focus on the concept rather than the clicks, allowing their mindset to stay in “strategic mode” rather than “functional mode”

for Developer

  • Removes end user need to be familiar with the tool due to natural language operation (easier adoption/switching tools)
  • Allows better understanding of customer usage and desired features based on recorded commands
  • Remarkable functionality worth users sharing ( social marketing)
  • Customers that switch CAD tools for this functionality can have a multiplier effect if their supply chain also switches

Market Size

The Global CAD industry is approximately $7 Billion dollars in size. Now that is the total market, and a significant portion of these are contract manufacturers, mostly using the CAD tools to open customer files for quoting and dimensioning purposes. These features are mostly not of much interest to that group.

However, the groups that would see a significant benefit from this would be research and development, industrial design, consumer products, aerospace, automotive, etc. These groups make up a significant portion of the total CAD market, though the data isn’t clear. While it appears, 80% of companies who own CAD tools are the manufacturers, due to their significantly lighter design loads, the 20% who are the the top of the supply chain, use significantly more CAD licenses per company, and as a result likely make up 80% of the license count. Since the numbers aren’t clear, with no unifying data source, let’s assume 50% of licenses in existence would benefit from these features.

Total license count in existence today isn’t clear. So, I’ll use some rough numbers for calculation. 50% of $7,000,000,000 is $3,500,000,000. SOLIDWORKS is middle market for CAD tools. The professional package which is the mid-level tool is $1,500 a year for subscription. So if we divide those two numbers $3,500,000,000/$1,500 = There is approximately 2.3 million CAD seats in use. That number seems to be in the right ballpark, SOLIDWORKS claims to have the largest install base and their numbers range from 600,000-1,000,000 user. Then there is CATIA, NX, Creo, Inventor, SOLIDEDGE and some smaller players making up the difference.

With 2.3 million users of CAD. If users see productivity gains of 10%, assuming an average burn rate of $80,000 per year, that’s $8,000 of value created from this feature. If that makes companies want to upgrade licenses for $500 to get this capability then the total addressable market size is 2.3 million * $500 = $1.15 Billion! Roughly 16.4% of the total existing CAD market.

If a single company could capture this portion of the market then there could be gains much larger than 16.4% found in revenue. So this idea represents a significant expansion of what is currently possible in the CAD market.

Competitive Advantage

  • Hard to replicate quickly due to complexity, 1st mover will be ahead of market.
  • First person to get users benefits from AI learning ahead of other companies and as a result, making it hard for competitors to catch up.
  • Catches people attention as a new revolution and lends itself to social sharing (network effect).

How to develop this idea further

With technical overhead rising due to compounding knowledge and technologies, it will be necessary to involve the potential customers throughout the entire development of this product, building a community around the idea early on. Otherwise sinking huge amounts of money and time into a product whose implementation ends in a form no one cares about is possible. Upfront customer interaction is both the prevention and the cure.

It’s tricky to navigate this path because the community needs to be all moving towards the same goal. It will need to be a focused group. With too many different needs represented, the final product will come out meaningless to most. For exactly this reason, a traditional, big company product development strategy will struggle to pull this off due to an early focus on big communities, which makes sorting priorities a high effort task, and a mostly meaningless endeavor.

Moving into the age of AI, the technical overhead is rising. The concepts behind AI require specific knowledge that lowers the amount of people that can do it compared to general programming. Starting an AI project may require an AI developer, a web developer/programmer for additional parts that aren’t linked specifically to the AI, and a significant investment in computing power. A well-rounded team for this type of project would include:

  • AI voice algorithm Developer
  • AI data Manager
  • IT Manager
  • Product Manager
  • Applications Engineer
  • Marketer/Sales

Dealing with the fear of AI

Artificial Intelligence (AI) is creating quite a reputation as something that should be feared, something that is going to take over all the jobs in the world. In the long-term, that will be something science and society has to figure out. In the short-term, it’s one of the technologies that’s actually going to free people to do the work of their lives. In this case, those people are designers, engineers, scientists, and anyone else utilizing CAD tools today. It will free them up to do more experimentation. To spend more time with ideas, or research, or serving customers.

The future will fracture with more products than ever, produced with groups of people in mind, rather than the masses. That’s because market has moved from the masses to the social. And only remarkable items make it. That and the idea that remarkable is in the eye of the beholder leads to the conclusion that there will be a need for more products than ever in the coming decades, and the people responsible for those designs will need more time freed up to think about them than ever before.

There is a lot for people to fear. There is also a lot to embrace. It may be the “job killer” but it could also be the “career enabler”. It’s all in how it’s marketed, created and adopted.

This idea has different value and implementation depending on the company that wants to investigate the opportunity and its impact on their product portfolio further. I’ve done a lot of thinking about the topic and what it would look like, feel like, how it would be implemented and marketed. If you want to explore the idea deeper together, I believe it could mean a growth of 20+% to any CAD company in the bottom line going forward due to increased product upgrades, capturing competitive accounts seeking the functionality, and shifting customers to the cloud.

I would like to work with you to develop a project and implementation plan. Contact me at to start the conversation and schedule a time to talk.

Thank you for your attention,

-Brandon Donnelly