Support

A letter from Eric Vaughan, CEO, IgniteTech and Khoros

Stability isn't boring. Stability is what lets you build.

Eric Vaughan

CEO, IgniteTech and Khoros

August 2026

There’s a specific kind of uncertainty that comes when a company doesn’t know what its future holds. Long-term planning stalls. Investment slows. Teams focus on maintaining what exists rather than building what’s possible.

It’s not anyone’s fault. It’s what happens when stability is missing.

Eight months ago that changed. Khoros found a permanent home with IgniteTech. No more transition, no more waiting, and a team finally able to focus on delivering for customers.

Here's what we built.
02

We'll start with the part that isn't flattering.

Like any platform with a long history, Khoros had accumulated technical debt. Support resolution averaged forty days. The engineering backlog ran to 930 items. AI capabilities existed but were barely used. Platform reliability needed work.

In a business where a social crisis unfolds in minutes, a forty-day support cycle isn’t a metric you defend. It’s one you fix.

So we got to work.

03

Support that actually supports.

We consolidated more than twenty request types — previously scattered across separate email addresses, internal systems and success teams — into one support portal with self-service automation behind it.

Then we changed how support works, not just how fast it runs.

40 → 17

Days, average resolution time

250 → 400

Tickets per week

9,000+

Tickets closed

5% → 60%

Resolved by AI

4,000+

Tickets solved without human intervention

70+

Intelligent automations built

We wrote 140 knowledge base articles to power those answers. The result: when your question is straightforward, you get an answer in seconds instead of weeks. When you need a human, our agents are available, because they aren’t buried in routine requests.

AI resolving 60% of tickets isn't replacing people. It's freeing them.
04

The backlog, the bugs and the uptime.

Over the last three months we’ve cut platform downtime by 97% and made incident recovery 81% faster.

Every capability we shipped was validated with customers before it went in.

930 → 169

Engineering backlog, 82% cleared

500+

Bugs resolved

270+

Security vulnerabilities closed

100+

High-priority accessibility issues fixed

05

Eight months, in one table.

Metric

Result

Support resolution time

40 → 17 days (60% faster)

Support throughput

250 → 400 tickets/week

AI resolution rate

5% → 60% (12×)

Tickets solved by AI

4,000+

Engineering backlog

930 → 169 (82% cleared)

Accessibility issues resolved

100+

Security issues resolved

270+

Platform downtime

97% reduction

Incident recovery speed

81% faster

Iris X integration build time

10 days

These aren’t targets. They’re results, measured between December 2025 and August 2026.

06

Three products, shipped.

Launched April 10, 2026

Aurora AI

Not an upgrade. A new platform, an AI-native reimagining of Khoros Community.

Launched April 10, 2026

Iris® AI

Khoros Care and Khoros SMM, rebuilt and merged into one platform for listening, publishing, engagement and analytics. Supports Facebook, X, Instagram, LinkedIn, WhatsApp, YouTube, Google Business Messages and SMS. Organic content, paid ads, direct messages and reviews in a single view.

Launched July 2026

CommunityGuard

Edge protection for Aurora AI communities, classifying visitors into allow, rate-limit and block.

Iris exists because of a ten-day fire drill. When a social media vendor ended our API access with no commercially reasonable terms available, we built and shipped a full replacement in under ten days, at no cost to customers, with no interruption to their operations. That’s the story that changed how we build everything else. Read it →

07

In build — forward-looking

What we're building now.

Aurora’s community intelligence work is built on a foundation we call the Community Language Model — an intelligence layer that learns from your specific community rather than the generic internet. Two capabilities sit on top of it.

Answer Assist

The unanswered-question problem, solved carefully. We could have built an AI that answers every question instantly. We didn’t, because premature AI answers erode trust and discourage the human participation that makes a community worth having. Answer Assist waits — typically 24 to 48 hours, configurable — and only responds if no qualified human has. Your members still help each other. The AI is the safety net, not the first responder.

The best AI answer is still second best to a good human answer. That's why we wait.

AI Moderation

Content moderation has always forced a bad choice: expensive human review that can’t scale, or automated filtering that misses context. Aurora’s moderation system combines both and shows its work — more accurate than legacy spam filtering, cheaper, and with explainable outcomes your team can audit and configure.

Enterprise controls

Iris ships with SSO via SAML, two-factor authentication, multi-tenant isolation, role-based access control, admin-approved onboarding and complete audit trails.

08

A permanent home.

All of this came from one change: stability.

We made hard decisions to optimise operations and focus resources where they mattered. Not to cut corners, but to build a base that could support sustained investment.

Khoros is no longer in transition. The best product work happens when teams aren’t wondering what comes next quarter — when there’s trust, and the freedom to think long-term.

Stability isn't the destination. It's the launchpad.

We’re proud of what we’ve built. We’re nowhere near done.

If you want to see how we build now, bring us a requirement. We killed the roadmap →