Pillar — Understands
Correspondence analytics
Correspondence analytics turns your chats, emails, tickets and documents into structured data. The platform extracts the fields you need with LLM rules, classifies and tags messages by meaning, computes metrics and gathers everything into a single analysis loop. You can forward correspondence to a dedicated address or upload files — and the output isn't scattered messages but analytics you can act on.
- Agent 0:02
Meridian Clinic, good afternoon. How can I help?
- Client 0:06
Hi, I'd like to book an appointment with a GP this week.
- Agent 0:11
Of course. There's a slot Thursday at 3:30 pm — does that work?
- Client 0:17
Yes, Thursday works. Please book it for me.
- Agent 0:21
Booked. I'll send a confirmation by SMS. Have a great day!
Key facts
- What it analyzes
- Chats, emails, tickets and documents — all your text correspondence in one loop.
- Data extraction
- LLM rules pull structured fields out of text: amounts, deadlines, IDs, contacts, statuses — whatever you need.
- Classification and tags
- Messages are sorted into categories and tagged by meaning, not by keyword matches.
- How to load it
- Forward emails to a dedicated address or upload files directly — no integrations required to start.
- Metrics
- Every message is scored with metrics and synced into analytics — dashboards and search across the whole archive.
All your correspondence in one loop
Scattered channels come together in one place, where the same analysis tools work over every piece of text.
- Chats, emails, tickets and documents are processed by the same rules — no manual sorting.
- Forward emails to a dedicated address and they enter the loop automatically.
- Upload files directly: you don't need to integrate anything to get a first result.
Structured data extraction
The platform pulls exactly the fields that matter to you out of free-form text — the rule is written by AI and runs deterministically.
- LLM rules extract amounts, deadlines, numbers, contacts and statuses right from the correspondence.
- Extracted values land in message fields — you can filter, count and export them.
- The rule is reusable: set it up once and it runs across your entire message flow.
Classification and tags by meaning
Messages are sorted into categories and tagged based on the meaning of the text, not on keyword matches.
- AI tags are assigned by meaning: 'complaint', 'urgent', 'billing question' — from what the message actually says.
- A hierarchical topic taxonomy reveals the structure of your entire message flow.
- Semantic search finds what you need 'by meaning, not words' — even when the wording differs.
Metrics and analytics over correspondence
Scored messages turn into numbers: you see trends, volumes and problem areas across the whole archive.
- System and custom metrics are computed on every message and synced into analytics.
- Dashboards show the dynamics of topics, deadlines and handling quality across the whole flow.
- 100% coverage of your correspondence, not a sample of a few read emails.
Product
Correspondence analytics
- 1
Connect your correspondence
Forward emails to a dedicated address or upload chats, tickets and documents as files.
- 2
Set up extraction and tags
Describe which fields to pull and how to tag — AI assembles the extraction and classification rules.
- 3
The platform scores the flow
Every message is automatically classified, tagged by meaning and enriched with extracted fields and metrics.
- 4
Analyze and decide
Review dashboards, search by meaning and export the data — correspondence becomes a source of decisions.
FAQ
Frequently asked questions
Which types of correspondence are supported?
Chats, emails, tickets and documents — all your text correspondence is processed in one loop under the same rules for extraction, classification and tagging.
How do I load correspondence into the system?
Two ways: forward emails to a dedicated inbound address, or upload files directly. You don't need to set up integrations to get a first result — the text itself is enough.
How does data extraction work?
You describe the fields you need, and AI generates an extraction rule — code that deterministically pulls amounts, deadlines, numbers, contacts and statuses out of the text. The values land in message fields, where you can filter, count and export them.
How are tags by meaning different from keyword search?
AI tags are assigned from the meaning of the text, not a word match: a complaint email is recognized as a complaint even if the word itself never appears. Semantic search works the same way — 'by meaning, not words'.
What do I get in the end?
Not scattered messages, but structured data: messages sorted into categories and tags, extracted fields and metrics. All of it is available in dashboards, filters and semantic search across your entire correspondence archive.
Related capabilities
Ready to start?
Turn every conversation into data, knowledge and action
Start by analyzing your conversations — no risk, no bots required. The platform turns your archive into data and a knowledge base, and voice agents plug in when you're ready.
Free plan, no card required.