AI Application Overview and Analysis
Introduction
Artificial intelligence (AI) continues to grow year-after-year. Therefore, the amount of applications leveraging AI have multiplied. This leaves users, like me, rather overwhelmed and unfamiliar with the possible applications/uses of AI.
Because of this, further expleration of is needed to shed light on the current AI world. This involves understanding what applications are avaliable, how are these applications used, and who is using these applications.
A user on kaggle, Muhammad Talha Awan, has a shared a collection of nearly 5,000 artificial intelligence tools along with descriptions, uses, charges, reviews, links, and overarching categories. A data dictionary is below, including examples.
| Variable | Description |
|---|---|
| ai_id | A primary key to keep track of each application. |
| AI Tool Name | The name of the AI application. |
| Description | A short summary of the application’s mission or goal. |
| Free/Paid/Other | The payment plan involved in using the application. |
| Useable For | Intended uses of the AI tool |
| Charges | The pricing structures associated with using the application (if applicable). |
| Review | Number of user-generated reviews the application has received. |
| Tool Link | The link to the homepage of the AI tool. |
| Major Category | Categorizes the AI tool into broader domains. |
| Price | Only the price involved in the charges variable. |
This data set will be used to determine what artificial intelligence tools are available for use and what their possibilities are. The upcoming analysis will also examine the cost pertaining to using these applications. Later on, an analysis of reviews on select AI Applications will be reviewed to contribute to the understanding of artificial intelligence.
AI Data Set Analysis
The following analysis will help describe the very large collection of nearly 5,000 artificial intelligence tools. With such a large amount of AI tools, this will give good grounding for the totality of AI tools.
Major Category Distribution
To get a good understanding of what applications are available for use, the distribution of the broad types of AI tools will be used for this analysis. This will come in the form of a bar plot.
As expected, the majority of applications come in the form of business. This is likely due to the wide variety of businesses with different needs. It is possible and maybe probable that companies engineering AI tools generate their money by angling their applications for the use of large enterprises with big wallets. This possibility will be discovered later in our analysis.
Text is the second most common AI application. This makes me curious about what differentiates AI text tools. I would infer that they act in similar ways. However, writing text is a very consuming task so it does seem reasonable to alleviate many of that time consuming task. There is much demand in this area.
Lastly, with how many AI applications categorized as “other”, I find it important to dig deeper into describing these applications. That is too large of of a group to leave uncategorized.
“Other” Category Breakdown
A break down the broad category of “other” is heavily needed. This will help shed light on the totality of tools available. Therefore, the 10 most common uses will be extracted, with the Used For variable, to determine what the AI tools in the “other” category are able to do.
The ‘Used For’ variable was broken down into individual instances and count the occurrences of each use for this analysis.
| Useable For | Frequency |
|---|---|
| life assistant | 238 |
| fun tools | 130 |
| research | 102 |
| resources | 79 |
| healthcare | 73 |
| search engine | 63 |
| travel | 54 |
| experiments | 49 |
| dating | 40 |
| gift ideas | 30 |
Through this table, the reasoning behind an “other” category is discovered. There is a large variation of AI tools that are hard to categorizes.
The most common AI application in the “other” category is a life assistant. This could have multiple meaning, but life assistant AI tools are built to answer to your every need, improve productivity and efficiency. They have the ability to send emails, plan out your day, schedule appointments, and much more.
Another interestingly AI tool with a large amount of applications is Healthcare. After further research into the descriptions of these applications, these AI applications have a multitude of abilities. Some of these abilities include medical imagining, pregnancy assistant, therapy, personalized medicine, and even drug discovery.
One of the more interesting discoveries on this list is the amount of dating AI tools. That is not something I would want to rely on a computer for, but people are entitled to their own choices. I could see a use for picking out date locations or activities, but most of these dating applications are either an AI partner or they exsist to answer dating questions.
An AI tool that helps with the details of traveling is something I would appreciate. That is something that is one of the more stressful and expensive. I would assume the demand for this would increase with more awareness.
AI Uses (Most Common and Least Common)
Most Common
Similarly, It is important to do this on a broad spectrum as well. This will allow for the realization of the most common uses and see what AI tools are frequently designed to accomplish. The same analysis done above will be utilized but with all applications.
| Useable For | Frequency |
|---|---|
| productivity | 445 |
| life assistant | 295 |
| general writing | 272 |
| developer tools | 271 |
| image generator | 249 |
| education assistant | 229 |
| social media assistant | 202 |
| fun tools | 200 |
| startup tools | 192 |
| copywriting | 191 |
Once again, life assistant near the top. Productivity, takes the highest slot for amount of AI applications following under that use. This is similar to the life assistant but targeted towards business. Some of these uses include resume/cover letter building, platforms for team communication and research, as well as text processing.
A group of AI tools seen in both lists was “fun tools”. This is another broad name. The description of these applications indicate that these AI tools are designed to generate poems, tell jokes, ask questions to historic philosophers, etc. These are applications many could use to pass time.
Social media and low-code applications are tools I believe many do or would appreciate. There are many companies, and especially sports teams, that have to send out a multitude of social media posts. I use to work in the communications department of a college athletic department and would have enjoyed having this around for my benefit. Low code applications are designed to write out code in multiple languages or critic previously written code. This could be used in both the business word, but also in education side of coding to aid the learning process.
Least Common
There is also an importance in understand which AI applications are the most unique. This could be due to the tools being designed for more niche markets or tools that are cutting-edge and the first to engineer an AI tool for a specific purpose.
This will be done by examine the least common occurring uses of AI tools.
| Useable For | Frequency |
|---|---|
| religion | 12 |
| logo generator | 20 |
| fashion | 21 |
| personalized videos | 21 |
| memory | 22 |
| paraphraser | 27 |
| fitness | 29 |
| real estate | 30 |
| gift ideas | 31 |
| presentations | 32 |
None of these findings were all that interesting. I expected some application to have sole ownership of a single use, however that was not the case. I do not seem many of these to be that revolutionary.
The one tool worth mentioning is religion. These applications are designed to be spiritual guides or help with sermons. Faith is a concept that is differs greatly from the way computers are designed and would struggle turning to a robot for guidance in this aspect. I believe religion should be more internal or self-reflecting.
Pricing by Category
Like mentioned before, pricing is important when users are examining possible AI tools. It was hypothesized that business AI tools would be relatively more expensive due to wealthy corporations.
The process of creating the box plot and summary statistics below involved converting all prices to an amount associated with one year of use. This allows for control between differing pricing plans (for example, some applications pay by month or week). For simplicity, applications that paid for individual use (for example, a text tool that costs $25/1,000 words) were excluded from this analysis.
| Major Category | Count | Mean | Median | SD | Min | Max |
|---|---|---|---|---|---|---|
| 3D | 40 | 350.67458 | 0.00 | 1461.9726 | 0 | 7188.00 |
| audio | 330 | 144.13659 | 47.94 | 343.8379 | 0 | 3588.00 |
| business | 1202 | 286.57393 | 29.00 | 1553.7831 | 0 | 30000.00 |
| code | 583 | 335.45199 | 23.88 | 2351.6905 | 0 | 48000.00 |
| image | 740 | 97.60784 | 10.00 | 264.7199 | 0 | 3599.88 |
| other | 849 | 183.53957 | 0.00 | 2887.0852 | 0 | 79164.00 |
| text | 1115 | 171.63018 | 59.88 | 392.0987 | 0 | 7188.00 |
| video | 110 | 193.89170 | 120.00 | 295.5463 | 0 | 2400.00 |
With the way that AI tools are set up, many of them being free, it is important to a look at both the mean and the median of each category. Extremely high priced tool with cause the mean to surpass the median by a fair amount. When looking at these results, there is a very large impact of applications that are usable with no payment. This makes gauging price different types of applications hard to determine. The pricing variations of the paid applications are of interest to me, so I will conduct this analysis again with the removal of any tools that are free fro use.
That said, there are still some takeaways to draw from this. Video AI tools are often priced higher more often than other applications. Non-video AI tools are relatively similar when it comes to median price. However, when examining the the average, or mean, price, 3D AI creation and codign tools exceed the rest of categories.
The highest priced application comes from a health care tool. The tool is JADBio and is described as “a no-code machine learning tool that automates the discovery of biomarkers, making it ideal for researchers in drug discovery, biomarker identification, and response to treatment studies.” These is a very impressive AI tool and has earned reasonable value.
The findings above also allow us to estimate that majority of tools that fall under the 3D and other categories are free.
Price Variations Among Paid Applications
This section removes all of the free applications generates the same table and plot as the previous analysis.
| Major Category | Count | Mean | Median | SD | Min | Max |
|---|---|---|---|---|---|---|
| 3D | 10 | 841.6190 | 108.00 | 2236.7508 | 1.2 | 7188.00 |
| audio | 180 | 217.8064 | 119.88 | 403.5499 | 0.0002 | 3588.00 |
| business | 602 | 458.4231 | 120.00 | 1945.6318 | 0.01 | 30000.00 |
| code | 266 | 562.4496 | 144.00 | 3026.3546 | 0.000035 | 48000.00 |
| image | 437 | 142.9497 | 59.88 | 310.1734 | 0.0015 | 3599.88 |
| other | 259 | 543.5322 | 59.00 | 4954.8863 | 0.0025 | 79164.00 |
| text | 672 | 254.3804 | 120.00 | 454.8579 | 0.0002 | 7188.00 |
| video | 71 | 240.3165 | 192.00 | 311.8407 | 0.3 | 2400.00 |
The table and box plot here are more meaningful in my opinion. It is likely that better applications have a price associated with them.
There are not many paid 3D AI tools available, both those that are relatively expensive in the AI tool market. The values in the other column are likely skewed due to expensive health care tools that dominated that category, this can be realized through the high standard deviation. Business tools are on the higher side, price wise, but they are not one of the top-three most expensive categories. The coding tools also have a large standard deviations, meaning that the pricing plans of these applications are not very similar.
AI Application Review
Now, it is time to understand who is using AI tools and what is the perception of these AI tools. To do this application reviews have been scraped from https://www.trustpilot.com/.
Ten Applications from our original data set have been carefully selected, based on their availability and review count. The reviews of these applications were scraped to add value to this quest for clarity. The AI tools selected and descriptions of those tools are listed below:
Applications
| Tool Name | Description |
|---|---|
| AdCreative.ai | Data-backed, AI-Generate ad creatives that outperform competitors. |
| Aiseo.ai | Supercharge your writing skills with AI-generated, Search Engine Optimized content. |
| Autowriter | Given a few words, AI will automatically create blog articles, product descriptions, artwork and more. |
| GetResponse | Automates email communication to grow contact list and send emails. |
| Kickresume | Creates beautiful resumes and cover letters with the help of artificial intelligence and customizable templates |
| Miro | A digital work space using native capabilities for business teams to build out visions. |
| PFP Maker | Generates hundreds of professional pictures with AI to showcase your best self online with the input of one photo. |
| Text Cortext AI | Catered to unique communication styles to advance writing and research capabilities. |
| Undetectable AI | Detects AI-written content and matches writing to the quality of human writings. |
| Wondr AI | Creates thousands of one-of-a-kind products with AI and posts products to destinations of your choosing. |
Review Analysis
The review information includes every review on trust pilot for these ten applications. The data is formatted so that every row includes the name of a reviewer, the location of the reviewer, the date of the review, the rating (scale 1-5), and the comment associated with the review. That data is leveraged to add to this continued understanding of the current AI environment.
Usage
It is important to examine how much use has increased within Artificial Intelligence. This is just a small sample of AI applications and is on the basis of reviews, but I am assuming that increased reviews is a result of increased use. The information found is hard to translate to all of AI but it will be able to estimate a relative pattern.
It is extremely evident that the amount of total reviews, among all of these ten applications, has increased heavily in the past 13 months. This is easy to assume as AI becomes more popular, however this helps generate some evidence to support that theory. I theorize that this will only increase in the future as AI uses increases, but at some point there will be a plateau as AI use moves closer to the numbers of computer use.
Location Popularity
The location at which reviews come from could have multiple meanings. The first being these countries consist of people that are more willing to leave a review. The second being that these AI applications are used more commonly in these countries. While it is impossible to control for the first meaning with this given data set, it has been chose to rely more strongly on the alternative reasoning.
| Ranking | Location | Total Reviews |
|---|---|---|
| 1 | United States | 1047 |
| 2 | India | 412 |
| 3 | Pakistan | 246 |
| 4 | Great Britain | 245 |
| 5 | Australia | 237 |
| 6 | Canada | 148 |
| 7 | Germany | 130 |
| 8 | Nigeria | 128 |
| 9 | Philippines | 122 |
| 10 | France | 120 |
Evidently, from the table above, the United States are well ahead in AI use. The United States are more about two-and-a-half times greater than India who is in second place, in terms of review totals. I was surprised to see that Nigeria is on this list due to the country being less technologically advance, but there is some reasoning due to English being their main language.
There are many issues with these conclusions as there are bias that are effecting the results as well. TrustPilot is largely an American website, therefore it is reasonable to assume that the American audiance is more active making reviews. It is also very possible that Americans are just more self-rigeous, leading them to make more reviews than people in other countries. This also does not take acount of American Companies making reviews from foreign subsidaries. It is also important to note that there is no weighting of a countries population in this ranking, it is strictly on review totals.
Location Rating Breakdown
The countries that heavily involved in reviewing these AI applications is know due to our previous analysis. Now, ratings differ by location becomes a question of meaning. The following displays the 10 boxplots that portray each of the top ten countries, listed above, in terms of how they rate products.
The median rating given out by every country for these ten applications is a 5, meaning at least half of the reviews coming out of each country is a 5. That is really good news for these AI companies, but I assume that is a similar finding for most companies.
The interesting findings come when looking at the lower quartile. The top row of boxplots does not identify a lower quartile, meaning that almost all of the ratings given out were five and anything not a rating of 5 was an outlier. However, when looking at the bottom row, these countries are more apt to lower review ratings for these ten applications. With the exception of Great Britian, their lower quartiles stems down to the rating of 4 and only ratings of 1 and 2 are deemed outliers. Thus, these countries are harsher critics of these ten AI tools. Furthermore, Great Britain is seemingly the toughest critic, as the lower quartile reaches the rating of 1, the lowest possible rating. It can be estimated that British people are judge AI tools more harshly.
Application Rating Breakdown
The applications selected should be described in some fashion. TrustPliot does a great job of depicting ratings on a individual basis, but does not do a great job of comparing applications or convaying deeper statistics like standard deviations. Therefore, I have decided to generate that for myself and raking the ten selected applications according to there average rating.
| Ranking | Application | Total Reviews | Average Rating | Standard Deviation |
|---|---|---|---|---|
| 1 | Autowriter AI | 42 | 4.857143 | 0.6466208 |
| 2 | PFP Maker | 18 | 4.777778 | 0.9428090 |
| 3 | Text Cortex AI | 494 | 4.767207 | 0.5978812 |
| 4 | AISEO | 809 | 4.676143 | 0.8682730 |
| 5 | Kickresume | 1267 | 4.555643 | 0.9299421 |
| 6 | Wondr AI | 70 | 4.442857 | 1.1246347 |
| 7 | AdCreative.ai | 1320 | 4.146970 | 1.4857136 |
| 8 | GetResponse | 465 | 3.681720 | 1.7646521 |
| 9 | Undetectable AI | 54 | 3.648148 | 1.8949527 |
| 10 | Miro | 48 | 3.437500 | 1.8324063 |
It is interesting to see how the ten different AI applications stack up against one another. The three AI writing tools (“Autowriter AI”, “Text Cortext AI”, and “AISEO”) are in the top four, in terms of average ratings. It can be estimate that writing AI tools are rated higher than other applications but with this little data, it is hard to make that generalization. It is also interesting that there is not a correlation between review total and rating average. I would have inferred that those with higher reviews would be larger, as they are likely more active companies.
AdCreative.ai has some interesting summary data. It is the most reviewed AI tool out of the ten selected, however its rating’s standard deviation ranks near the bottom. With higher review counts, I would suspect the standard deviation to be lower, but this really shows that those review this ad creator have mixed feelings about the product.
It is important to note that some of the applications have much higher review totals and therefore, the average rating is likely to closer to the real review ratings of the populations. Also, I believe that it is less likely for an application to receive a review if that rating is anything but a 1 or a 5. This is due to strong feelings either way provoking a review. This is just a theory, no research back this statement.
Application Popularity by Country
Lastly, I think it is important to shed light on what applications are more popular in different areas. Our list from earlier of the highest reviewing countries is a rather diverse list, so the application they review the most will be determined. This will for a solid educated guess on what application they use most. The following table lists the Location (country), Application (highest reviewed application), Reviews (the amount of reviews that country offered to that application, and Review Percentage (that applications review total as a percent of a country’s total reviews).
| Location | Application | Reviews | Review Percentage |
|---|---|---|---|
| Nigeria | AdCreative.ai | 82 | 64.06250 |
| Australia | Kickresume | 126 | 53.16456 |
| France | Kickresume | 63 | 52.50000 |
| Philippines | AISEO | 62 | 50.81967 |
| Germany | Kickresume | 64 | 49.23077 |
| Great Britain | AdCreative.ai | 107 | 43.67347 |
| Pakistan | Text Cortex AI | 90 | 36.58537 |
| Canada | AdCreative.ai | 47 | 31.75676 |
| India | Kickresume | 126 | 30.58252 |
| United States | AdCreative.ai | 261 | 24.92837 |
These finding are quite interesting. Nigeria, Australia, France, and the Philippines all have an application that makes up more than 50% of their overall reviews. It can be infer that Nigeria uses AdCreative.ai as a large part of their marketing industry. It is possible that this is the most popular AI tool in that country.
AEISO being the most popular AI application to be reviewed in the Philippines is rather intriguing. AEISO’s total reviews was high, but not to the extent of Kickresume or AdCreative.ai. Therefore, this is an unusually high percentage of reviews targeted at AEISO. This leads to further curiosity into why the Philippines is drawn to this writing tool.
It makes logical sense that the countries with the highest total reviews, United States and India, have less in one basket than other countries that review these applications less. This is likely due to a broader exposure to AI tools and higher usage in multiple industries.
Review Sentiment Analysis
There are multiple different AI tools that are designed to help in the writing process. Two that are rather large in this sector of AI are “Aiseo” and “Text Cortext AI”. Lets compare what the sentiment is across these two platforms by using reviews on https://www.trustpilot.com/. This will help derive some evidence to which writing AI tool is preferred or recommended.
To control for differing quantities of reviews, the 450 most recent reviews have been scraped to leverage our analysis. A breakdown of the two sets of reviews is listed below for reference.
| Ranking | Application | Total Reviews | Average Rating | Standard Deviation |
|---|---|---|---|---|
| 1 | Text Cortex AI | 450 | 4.757778 | 0.6164013 |
| 2 | AISEO | 450 | 4.657778 | 0.9019871 |
This analysis will dive into the words used when providing a review comment. This will help us better understand the perception of each application further than just a review rating. While each of us are entitled to shape our own opinion. The hope is to decide which of these writing applications is more recommended.
Positive vs. Negative
It is important to understand what scorable words, meaning words that are either negative or positive, are within these reviews. To do so words used at least 20 times in the totality of reviews for each application will be extracted. The values in the graph that appear negative are just demonstrating that it is a negative. For example, if “bad” appeared 20 times, its value would be -20.
These results are interesting. The first realization that comes to mind is that there are multiple words that show up 22 or 23 times Aiseo. This could be due to the presence of fake reviews as they are describing the application in the same ways. This is ironic because AI writing tools are being used for this analysis. The word “helpful” way seen 134 times over these 450 reviews, which is a great deal larger than the second highest use of a word and a very solid word a AI writing application would want to be described as.
There are very few words that appear more than 20 times for Text Cortext AI. This is probably due to a larger variance in review comments making it more likely to be real reviews.
The word “easy” is used extensively more both application. That is a common word I use when describing Artificial Intelligence tools, so this is a good sign for both applications.
There do not seem to be any words out of the ordinary for a AI writing tool application. These are the words that will have the most impact on sentiment as they are seen most oftne. Therefore, I feel comfortable moving forward without the removale any words.
Positivity by Month
AI Writing Applications are likely treated with more detail during specific times of the year. In the perspective of a student, an application like these can become very valuable during the end of each semester or quarter when there is a higher influx of essays. For business, on the other hand, there is likely a higher need during times with a large volume of press releases. This could happen at the end of each quarter or if there is something large going on in a political sphere.
Due to the reviews available. Only reviews between June and November will be on the following graph. None of the 450 last reviews occurred prior to June or in the month of December for Aesio.
It does seem that there is an increase in review positivity in the months of September and. November for Aiseo and Text Cortext AI joins in on the positivity increase in the month of November. This seems to fit our earlier predictions as midterms would occur around September and the semester begins to wrap up in November. Furthermore, many companies have their quarterly reports in the months of September and late November.
It is important to keep in mind that that the increasing positivity score above could also be the result of the increase of application use stated previously. After all, artificial intelligence is a rather new. This could also be due to the applications having flaws early on in their existence.
Sentiment Scores
Lastly, lets take a look at the sentiments used in these reviews. This will categorize scorable words into emotions. This will be done on the basis of the NRC Word-Emotion Association Lexicon and more information can be found here.
The number of words that fit each emotion between the two applications will be compared. This will help us understand, in further depth, how reviewers feel about these applications.
Right away, the biggest take away from the graphic above is that there are more words associated with nearly all emotions coming from the Aiseo reviews compared to the Text Cortext AI reviews have. This gives further evidence to Aiseo reviews being more word or scorable as the number of reviews was controlled for.
There is a great deal more positive words for Aiseo than Text Cortext AI, however, there is nearly the same difference in negative words. This allows us to consider that they could be comparably similar at equal word levels.
It is a good sign that the highest emotion, outside of positivity is with trust. One of the uncertainties that many have with artificial intelligence is that they cannot trust the work produced or the modeling behind it. This is probably the most meaningful emotion when evaluating AI writing applications. Aiseo wins this category by a sizable mark, but Text Cortext AI has a high score as well.
One of the more interesting findings comes when examining the word counts that align with fear. This is the only category that Text Cortext AI surpasses Aiseo in word count. This is not a emotion I would want to be associated with an AI tool. This emotion total for Text Cortext AI would most likely be higher when the number of reviews is weighted.
Recomendation
Both of these application are rated rather highly and the review content seems to be overwhelming positive. The reviews for Aiseo happen more often and relatively more wordy than reviews for Text Cortext AI. There was also no control for fake reviews. This is important limitations and biases that had the ability to eftected our analysis.
With all of that said, it seems as if Aiseo is the more recommend AI writing tool. The reviews for Aiseo were relatively more positive and trusting. It overwhelmed Text Cortext AI in nearly every aspect and, without ever using either application personally, I would recommend Aiseo between these two writing tools.
Conclusion/Implications
Overall, this discussion has been able to spread some light on the current scene of Artificial intelligence tools. The types of AI tools available were identified, allowing for potential users to find, what they deem useful. The pricing of said applications was detailed in order to spread some awareness ahead of an AI tool search. Lastly, reviews of ten AI tools were leveraged to determine the perceptions of those different tools and to make estimations about the whole AI world.
There was a rather small portion of the total review data available on AI Applications, however it is a meaningful amount that was deemed accessible. Many of the review findings cannot be generalized as they involve just ten AI applications and reviews from just one site. That said, there is still meaning to be found in the information provide above. Many estimations can be developed to follow the analysis above even with the accuracy is at question. All in all, this was an opportunity to dive deeper into the new artificial world we find ourselves in and provide some clarity to the evolving sphere of artificial intelligence.