Chatbots / Assistants
We build chatbots around what each client actually needs, and they tend to fall into two kinds. User-facing ones sit on your site or app and answer your customers from your own content. Internal ones work behind the scenes and help your team make sense of the company's own files. Here is what each does.
User-facing chatbots
The assistant your visitors talk to. It answers their questions from the content you already have, your docs, your FAQs, your site, and only from what it can actually find, so replies stay accurate instead of invented.
IronFox Assistant
Online
Retrieving from your sources
- Understand the question
- Decide how to answera real question
- Search your contentdocs + FAQs
- Keep the best matches3 passages
- Write the reply
How it works
- 1
Read the question
The assistant works out what the visitor is actually asking.
- 2
Decide how to answer
A quick hello or thank-you it handles on its own; a real question it takes to your content.
- 3
Search your content
It looks across your docs, FAQs, and site for the passages that fit the question.
- 4
Keep the best matches
Only the passages that genuinely answer the question make it through.
- 5
Write the reply
It composes an answer from what it found, so it stays grounded and never guesses.
IronFox Assistant
Online
Retrieving from your sources
- Understand the question
- Decide how to answera real question
- Search your contentdocs + FAQs
- Keep the best matches3 passages
- Write the reply
How it works
- 1
Read the question
The assistant works out what the visitor is actually asking.
- 2
Decide how to answer
A quick hello or thank-you it handles on its own; a real question it takes to your content.
- 3
Search your content
It looks across your docs, FAQs, and site for the passages that fit the question.
- 4
Keep the best matches
Only the passages that genuinely answer the question make it through.
- 5
Write the reply
It composes an answer from what it found, so it stays grounded and never guesses.
Internal chatbots
An assistant for your team, not your customers. It works over the company's own files, from spreadsheets and reports to raw figures, and helps staff find, summarise, and actually analyse what is inside them.
IronFox Assistant
Online
Working through your files
- Understand the request
- Find the right filessales-q4.xlsx
- Write the query
SELECT client, SUM(revenue) AS total FROM sales_q4 GROUP BY client ORDER BY total DESC
- Run it over the data1,284 rows
- Return and save the resulttop-clients-q4.csv
Saved totop-clients-q4.csv
How it works
- 1
Read the request
It works out what the employee is asking of the data.
- 2
Find the right files
It locates the relevant company files, from documents to spreadsheets full of numbers.
- 3
Decide the approach
A straight lookup for a simple question, or real analysis when the figures need crunching.
- 4
Write and run the query
For anything involved it writes the code to group, rank, and compute across large files, then runs it.
- 5
Answer and save
It returns the result on screen and saves it as a file, so the team can reopen it without redoing the work.
IronFox Assistant
Online
Working through your files
- Understand the request
- Find the right filessales-q4.xlsx
- Write the query
SELECT client, SUM(revenue) AS total FROM sales_q4 GROUP BY client ORDER BY total DESC
- Run it over the data1,284 rows
- Return and save the resulttop-clients-q4.csv
Saved totop-clients-q4.csv
How it works
- 1
Read the request
It works out what the employee is asking of the data.
- 2
Find the right files
It locates the relevant company files, from documents to spreadsheets full of numbers.
- 3
Decide the approach
A straight lookup for a simple question, or real analysis when the figures need crunching.
- 4
Write and run the query
For anything involved it writes the code to group, rank, and compute across large files, then runs it.
- 5
Answer and save
It returns the result on screen and saves it as a file, so the team can reopen it without redoing the work.