Frequently asked questions
Can I log email activity in HubSpot without storing the message content?
Yes, this is the core of what Email Classifier does. It reads each message in flight to classify it, then logs a native email engagement on the HubSpot timeline. The raw email body is never stored. What we keep is the classification, a short list of extracted facts, a subject preview of up to 200 characters, and standard metadata, and even those expire after 90 days.
HubSpot only logs replies to threads we already logged. Can I capture all inbound email automatically?
Yes. Because it connects to Microsoft 365 through Microsoft Graph, it sees inbound email to your monitored mailboxes whether or not anyone logged the thread, and records the activity automatically. There is no forwarding, no BCC address, and no per-thread logging to remember.
I only want to log email to and from existing CRM contacts, not create noise. Can I control that?
Yes. It matches against your existing contacts and companies, and by default it keeps personal and free webmail only correspondence out of the CRM. You choose which mailboxes and domains are monitored. It can create contacts from genuine business correspondence, or you can keep it to your known contacts.
Can it categorise or classify incoming emails by type?
Yes. Every message is classified against a preset built for your industry, with a request type, the sender relationship, and signals such as urgency or buying intent. Your timeline then shows what each email was about, not just that one happened.
We worry about people oversharing sensitive information into the database. Who can see the email content?
No message content is written into HubSpot. Only the classification and a short subject preview land on the record, so sensitive detail from the body never enters your CRM for staff to browse. The content is read only in flight to classify it, then discarded immediately.
Does it enrich contacts and companies?
Yes. It keeps last contacted current, pulls roles and titles from email signatures, creates contacts from real correspondence, and surfaces recent activity and signals on contact, company and deal cards. It never writes to your deals.
Do we need to install an add-in on every Outlook or Gmail?
No. It connects once at the tenant level through a Microsoft 365 admin consent. There is no add-in to deploy on staff devices and nothing for your users to install.
Where is our data processed, and is it suitable for regulated industries?
All processing and storage takes place in Australia, in AWS Sydney, with model inference on Amazon Bedrock in the Australian region group. Nothing leaves the country. Because content is never stored and derived records expire after 90 days, it is a defensible answer for IT, privacy and procurement reviews.
What exactly do you store, and for how long?
The classification result, a short list of extracted facts, a subject preview of up to 200 characters, and metadata such as sender, recipients, subject, date and message id. Never the body. Everything carries a 90 day expiry and is deleted automatically.
What does it cost?
Ten dollars for 35,000 classified emails. One credit classifies one email, you start with 2000 free, and you can top up or turn on auto top-up any time. There are no per-seat fees and no lock-in.
Still have a question? Email support@redyellowblue.com.au.