Doc & Writer Book a call

Your knowledge base, kept right.

We run knowledge operations for B2B software teams. Each month we test what your help center and AI agent tell customers, fix what’s wrong, and keep content current with each release.

Help center / Billing / Exporting invoices
Status
Out of dateVerified
Owner
NoneBilling team
Last reviewed
14 months agoThis week

Exporting invoices

Go to Billing, choose Export, and pick CSV or PDF. All plans can export up to 24 months of invoices.

Go to Billing, choose Export, and pick CSV or PDF PDF. CSV export is on the Growth plan and above. All plans Starter and Growth plans can export up to 24 12 months of invoices.

Sample article. The company and details are illustrative.

Results from programs our team has led

96%Case accuracy across 15+ markets, up from 88%. Meta.
90%Fewer client escalations within 60 days. Meta enterprise support.
5xKnowledge base adoption after a migration for 10,000+ users. O-I Glass.
12Product lines migrated to Adobe Experience Manager Guides. Hunter Douglas.

Help content goes stale with every release.

Your support team learns which articles to skip. Your AI agent repeats them to customers, and some agents count those conversations as resolved.

We handle conversations a month. About % of answers are wrong, % of those become tickets, and each ticket costs us $.

$216,000a year in tickets caused by wrong answers, about 1,500 a month.

Example values. Edit any number. Gartner put the average live-channel contact at $8.01 in 2019, and Intercom counts a conversation as resolved when the customer leaves without asking for more help. Sources: Gartner, Intercom.

Knowledge Ops

A monthly service that does the work of an in-house knowledge team, without the hire.

TestFixPublishReviewReport Every month after a month-one audit
  • Accuracy testingReal customer questions, run against your help center and AI agent each month.
  • Fixes ranked by costThe most expensive wrong answers get fixed first.
  • Release contentArticles written and approved before each release ships.
  • OwnershipA named owner and a review date on each article.
  • AI-ready structureArticles split and titled so your AI agent retrieves the right one.
  • Monthly reportAccuracy, open gaps, and the tickets wrong answers caused.

From $7,500 a monthThree-month minimum. Month one is a full audit.

See what’s included

Automations run between reviews.

Review reminders, release triggers, and feedback routing, set up in Jira, Confluence, Salesforce Knowledge, SharePoint, or Power Automate.

  1. Review reminder sent to the Billing owner for 3 articles
  2. Release ticket PROD-482 opened a documentation task for Saved Views
  3. A support agent flagged “Refund window” as wrong. Routed to Support Ops, due Oct 12
  4. “Refund window” corrected and approved
  5. Accuracy run: 188 of 200 answers correct. 12 failures reviewed
  6. Monthly report sent. Review call booked
  7. Sample activity from one month of Knowledge Ops. Names and numbers are illustrative.

Symptoms we fix

Each one has a probable cause and a remedy we build into Knowledge Ops.

Your AI agent gives confident wrong answers.

Probable cause

It answers from articles written before your last few releases.

Remedy

A monthly accuracy run, with fixes ranked by what each wrong answer costs.

Agents give five answers to the same question.

Probable cause

Duplicate articles, and no one owns any of them.

Remedy

One article per question, each with a named owner and a review date.

Help articles ship after the feature does.

Probable cause

Documentation sits outside the release process.

Remedy

A release trigger in Jira that opens a documentation task the day the release ticket does.

No one can say whether the help center works.

Probable cause

Page views are the only measure.

Remedy

A monthly report on accuracy, open gaps, and the tickets wrong answers caused.

Your knowledge base is moving to a new platform.

Probable cause

Migrations copy outdated content along with the good.

Remedy

An audit before the move, then a verified cutover. Our team has run migrations with zero downtime.

What colleagues say about our founder

Alexander Dade leads our client work. These come from his managers and peers at Meta and O-I.

…he owned the knowledge foundation behind it, driving a large-scale KB feedback audit and delivering the knowledge-base overhaul that made the Co-Pilot recommendations trustworthy and deployable as the program scaled from experiments to production.
Sharmilla AbhilashGlobal Operations Senior Program Manager, Meta
He effectively gathered information from various parties and distilled complex topics into easy to understand documents which were used across continents.
Michael HuTechnology Licensing Lead, O-I
His collaborative approach was essential to expedite the adoption of CCMS and LMS programs that were new for the glass manufacturing environment.
Javier Reyes, MSc, PMPProject Planning Director, O-I
One of Alexander’s standout strengths is his mastery of timelines and dependencies. He orchestrates multiple moving parts with precision, ensuring that projects stay on track and stakeholders remain aligned.
Jaime GallegosProgram Manager, Meta
I had the pleasure of managing Alexander directly for 8 months and was more than pleased with his skill and thought leadership.
Dalton ManciAI Learning & Knowledge Technology Architect, Meta

Read all 17 recommendations on LinkedIn

How it starts

Three steps, and no commitment until the second.

  1. Free assessment

    Send us your help center address. Within five business days you see 25 real customer questions, which ones your content answers wrong, and why.

  2. Month one: the audit

    We build a 200-question test set from your tickets, score your content, price the wrong answers, and assign an owner to each article.

  3. Every month after

    We fix what costs you the most, cover each release, re-test, and report.

Our guarantee

You approve the test set and the scoring rules. If accuracy has not reached the target we agree on by day 90, we keep working at no charge until it does. Read the conditions.

See what your help center gets wrong.

A free assessment of 25 real customer questions, with the article behind each wrong answer. Delivered within five business days.