Introduction

(Un)Perplexed Spready is an AI-driven spreadsheet software designed with purpose to utilize AI models for useful work in spreadsheets.

in this article, we are going to demonstrate how (Un)Perplexed Spready can be used in a real business use case, where artificial intelligence can combine reasoning with deep internet research in order to retrieve and compile business critical information.

We will also compare results retrieved by using different Perplexity AI models, namely: sonar, sonar-pro and sonar-deep-search.

Task Description

We have a dataset containing 2355 records of deals made by various VC firms. The dataset contains columns with company name, VC firm name, VC firm website, deal type, deal size and deal date. Our task is to find a person (investment lead partner) who made each deal for a VC firm.

In order to get such information, we would need to browse and compile data from multiple internet resources - often this is signaled in a blog post, in a social media post, on the partner's bio page, etc. Obviously, it would be a very cumbersome and tedious job if being done manually. Instead, we are going to hire Perplexity AI inside a spreadsheet to do this job for us.

As Always, our moto is: Let AI do the hard job, while you are drinking your coffee!

Our Tool - (Un)Perplexed Spready

(Un)Perplexed Spready is an advanced spreadsheet software developed by Matasoft, designed to bring artificial intelligence directly into everyday spreadsheet tasks. Unlike traditional tools like Excel or Google Sheets, it integrates powerful AI language models into spreadsheet formulas, enabling automation of complex data processing tasks such as extraction, classification, annotation, and categorization.

Key features include:

• AI-embedded formulas: Users can utilize advanced language models directly in spreadsheet cells to perform data analysis, generate insights, and automate repetitive tasks.
  Multiple AI integration options:
      PERPLEXITY functions connect to commercial-grade AI via Perplexity AI’s API, offering high-quality, context-aware responses.
      ASK_LOCAL functions allow users to run free, private AI models (like Mistral, DeepSeek, Llama2, Gemma3, OLMo2 etc.) locally through the Ollama platform, providing privacy and eliminating API fees.
      ASK_REMOTE functions enable access to remotely hosted AI models, mainly for demo and testing purposes.
• User-friendly interface: The software maintains familiar spreadsheet features such as cell formatting, copy/paste, and advanced sorting, while adding innovative AI-driven interactivity.
• Automated data management: Tasks like extracting information from unstructured data, converting units, standardizing formats, and categorizing entries can be handled automatically by the AI, reducing manual effort and errors.
• Flexible for all users: It is suitable for both beginners and advanced users, offering significant time savings and improved accuracy in data processing.
• By embedding AI directly into spreadsheet workflows, (Un)Perplexed Spready transforms the way users interact with data, allowing them to focus on creative and strategic work while the AI manages the labor-intensive aspects of spreadsheet management.

Perplexity AI

Perplexity AI is an advanced AI-powered search engine and conversational assistant that provides direct, well-organized answers instead of just links. It uses real-time web search and large language models to quickly synthesize information and present it in a clear, conversational format.

Key features include conversational search, real-time information gathering, source transparency, and support for text, images, and code. It’s accessible via web, mobile apps, and browser extensions, and is designed to be user-friendly and ad-free.

Perplexity AI is used for research, content creation, data analysis, and more, helping users find accurate information efficiently and interactively.

Methodology and AI-driven Formula Used

The full dataset contains 2355 records, but for the sake of this presentation, we are going to calculate only first 50 records.

We will be using PERPLEXITY3 formula, which means that we will pass 3 input ranges and 1 prompt as parameters in the formula.

1st input range will be column A ("Company"), 2nd input range will be column B ("VC Firm"), while for the 3rd input range, we will create a new concatenated column "Input 3" by using the following formula to concatenate multiple column into one input column: ="VC Firm Website: "&IF(C2="","unknown",C2)&"; Deal Type: "&IF(E2="","unknown",E2)&"; Deal Size: "&IF(F2="","unknown",F2)&"; Deal Date: "&IF(G2="","unknown",TEXT(G2,"dd.mm.yyyy"))

Now we can use Column A ("Company") as Input1, Column B ("VC Firm") as Input2 and newly created concatenated column I2 ("Input3") as Input3 in the formula of column "Lead Partner":

=PERPLEXITY3(A2,B2,I2,"For a VC venture deal, we know the name of the company which is given in the Input1, the VC firm which is given in the Input2, and other deal details given in Input3 such as the VC firm web site, deal date, the amount of the deal, and the deal type. By using all these informations, you need to identify which investment lead partner at the VC firm led the deal for the VC firm. For this you will need thoroughly investigate internet resources - often this is signaled in a blog post, in a social media post, on the partner's bio page, etc. Deeply search the web - including the company's website, the VC firm's website, the partners' LinkedIn pages, and the partners' Twitter pages. Return determined investment lead partner of the VC firm.")

Comparing Results

Results by using Perplexity AI sonar model

Results by using Perplexity AI sonar-pro model

Results by using Perplexity AI sonar-deep-search model

Comparison

We have put results of all 3 models into the same sheet and also added another column in which we combined results by using locally installed Ollama model gemma3n:e4b.

=ASK_LOCAL3(D2,E2,F2,"Input1, Input2 and Input3 are results of the same query, retrieved by 3 different LLM models, about a person who is the investment lead partner of a VC firm closing a particular deal. Analyze and combine information from Input1, Input2 and Input3 and return final conclusion about the lead partner. Return just the name of that person as the final result. If in doubt due to conflicting inputs, Input3 has prevalence over Input 1 and Input2.")

Let's compare the results:

We can see that the three Perplexity models (sonar, sonar-pro and sonar-deep-search) return considerably different results. While sonar-pro is most concise and straightforward, it failed to find the lead partner for many records. On other hand, sonar seems to be more complete but also pretty verbose, which is annoying if we just want concise final result.

As expected, the Sonar Deep Search returned the most reliable and most complete results, sometimes too verbose though.

But, there is a huge caveat with sonar-deep-search - it is very expensive! We measured price per formula calculation, i.e. price per cell and difference is staggering: 

Price per cell for sonar: 0.01 $

Price per cell for sonar-pro: 0.02 $

Price per cell for sonar-deep-search: 0.60 $

As you can see, sonar-deep-search is 60 time more costly than sonar, for exactly the same formula! For big number of rows the price difference can be huge, so before deciding to go with Perplexity AI deep search, you should first think twice, whether results quality justifies such huge price difference or not.

Additionally, while execution of sonar and sonar-pro for 50 rows took 1.5 to 2 minutes, for the same number of rows being processed with sonar-deep-search it took 1.5 hours! Not only that the price difference is huge, the runtime difference is also very big.

At the moment of writing this article, there was also an additional restriction present - while for sonar and sonar-pro there is no daily limit, for the sonar-deep-search there is a daily limit of 500 queries. You should check and take this into account too.

Conslusion

Perplexity AI is great tool for combining advanced LLM models with web search features, thus being the leader of AI-driven web search market. By using Perplexity AI, (Un)Perplexed Spready can be a perfect tool for compiling business datasets where information retrieval and deep internet research is required.

Additionally, (Un)Perplexed Spready enables you to combine results retrieved by different Perplexity models, utilizing free locally installed Ollama models, such as gemma3n, thus you don't need to spend extra money on that final comparison and compilation operation.

Summarizing everything, our final recommendation:

1. If you want to perform cost-efficient and fast web search and data compilation, go with Perplexity AI sonar and sonar-pro and combine their results using locally installed gemma3n model.

2. If you don't care about costs and execution time, and just want complete and straight-forward information, go with Perplexity AI sonar-deep-search.

3. If you don't care about costs, neither execution time and want to have maximum quality results, go with all three models (sonar, sonar-pro and sonar-deep-search), then combine their results using locally installed gemma3n model.

 

AI-driven Spreadsheet Processing Services

We are also offering AI-driven spreadsheet processing services with (Un)Perplexed Spready software.

If you need massive data extraction, data categorization, data classification, data annotation or data labeling, check-out our corresponding services here: AI-driven Spreadsheet Processing Services

Further Reading

Download (Un)Perplexed Spready

Purchase License for (Un)Perplexed Spready

Using (Un)Perplexed Spready

Leveraging AI on Low-Spec Computers: A Guide to Ollama Models for (Un)Perplexed Spready