September 2026

28-Sep-2026 — Brand voice

Choose the tone of the standard messages your customers receive — the welcome, the messages sent when a conversation is resolved, and the messages sent when it is handed to a person — per brand, from AI Configuration.

• Four tones — Friendly (the default, today’s wording), Professional, Matter-of-fact and Warm. Each shows a sample before you choose it, and View all messages shows the full set.
• Every standard message, in one place — grouped into Welcome, Resolution and Hand-overs, with when each is sent and whether it goes on tickets, chat or both. Hand-overs are grouped by what the customer needs to hear, so one message covers several reasons; your agents still see the precise reason on the ticket.
• Use your own words — edit any message, up to 500 characters. Revert goes back to the standard wording of the tone selected now. Changing the tone changes every message you have not edited and keeps your wording for the ones you have; the page says how many before you confirm.
• Disclosures stay put — the short notice at the end of an automatic message, saying it was sent automatically, follows the tone but cannot be edited, so customers can always tell when they are talking to the AI.
• Set per brand — the tone chosen for all brands applies to any brand that has not chosen its own.

Docs: AI Configuration — Brand voice

28-Sep-2026 — Documents

You can now upload files for the AI to answer from. Policies, runbooks and product documentation that live in a PDF, a Word document, Markdown or a text file no longer need to be turned into help center articles first.

• Four file types — PDF, Word (DOCX), Markdown and plain text. Up to 25 documents per brand, each up to 30 MB, and up to ten files at a time.
• Always internal — the AI uses a document to answer, but a reply to a customer never names it, cites it or links to it. Agents see it once in the ticket’s knowledge panel, marked (internal).
• Tables keep their headings — each row of a table is read together with its column headings, so a price or a specification that lives in a table can be found and answered from.
• Personal information removed — email addresses and phone numbers are removed as a document is added, following your brand’s Scrub personal information from internal articles setting. A document therefore cannot be used to give a customer a contact detail; put those in a public help center article instead.
• Text only — the AI reads body text and tables. Text that appears only inside images, such as a scanned page, is not read, and a document with several columns or a complex layout may be read imperfectly.
• Clear results — each file shows Processing, then Ready, or Failed with the reason in words, for example that the file is password protected or has no selectable text. A file with the same text as one already added is refused, and the message names the original. The file you upload is not kept once it has been processed.
• Rename, replace and delete — Replace swaps in a new version all at once: the current version keeps answering until the new one is ready. Delete removes the document and everything the AI learned from it.
• Who can change them — Admins and Knowledge Managers. Read Only users see the list with every control switched off.

Docs: Knowledge Sources — Documents

24-Sep-2026 — Website Content

Flexivity AI can now answer from a public website as well as from your help desk’s knowledge base. Connect your documentation site, and its pages are read into the same knowledge the AI uses, and cited to customers the same way.

• Knowledge Sources — a new section in the admin console, per brand, that lists everything the AI knows: your Zendesk or osTicket knowledge base and any websites you connect. The help desk knowledge base controls now live here, with article counts by bucket and a sync page.
• Whole site or specific pages — start from an address or a sitemap and follow links within the site, or list the exact pages to read. Optional path filters narrow what is read.
• Polite and safe — the crawler follows your site’s robots.txt, reads only public pages, and never removes pages unless it has seen the whole site, so a failed crawl leaves what the AI knows untouched.
• You choose when it refreshes — every week, every two weeks (the default), every month, or only when you click Sync now.
• Sync history — a log of every crawl with its outcome, the pages it indexed, and a short reason when one fails.

Docs: Knowledge Sources

22-Sep-2026 — Progressive Autonomy Launch

Flexivity AI now offers progressive autonomy — a capability ladder that lets you control how much the AI does on its own, from read-only summaries all the way up to autonomous ticket handling.

• Autonomous ticket handling — Flexivity AI can now reply to customers, ask clarifying questions, and resolve conversations without agent involvement, within bounds you configure per intent type.
• Seven-rung capability ladder — choose the right level of AI involvement for each type of ticket: Assist, Auto-Classify, Auto-Reply (single-shot), Auto-Clarify, Auto-Resolve, and beyond.
• Configurable per intent type — set different autonomy levels for different kinds of tickets using classification rules and attribute predicates.
• Built-in safeguards — every AI-generated reply is verified against your knowledge base before it reaches the customer; unverified claims trigger an immediate hand-over to a human agent.

Correction (25-Sep-2026): two statements in this note were inaccurate.

• Level names. The levels are Ignore, Label, Gated, Assist, Reply, Clarify and Resolve. “Auto-Classify” and “Auto-Reply” are not level names.
• Verification. Every claim in a reply must cite an article your knowledge base returned, and a reply with an uncited claim is withheld. That check is always on. The deeper check, Answer verification, where a second AI call judges whether each cited article actually supports the claim, is off until you turn it on, separately for tickets and for chat. See Answer verification.

Docs: Product Overview — Capabilities