Most impactful technical writing items

I used AI to save 99% of time

Atelio of FIS Global (Seattle) -- I used AI to read Slack messages and create a 1st draft of a developer guide in 1% of the time

Before (Sep 2025)	After (Sep 2025)
PROBLEM: Over two weeks of multiple daily group Slack messages, ten fintech PMs and SMEs were discussing and gradually deciding the specs, functionalities, details, and limitations of a new mobile app financial tool. After they gradually solved or finalized different aspects of the new product, they then asked me to write a developer guide based on their dozens of decisions.	MY SOLUTION: The Promptless AI tool reviewed all those conversations and listed only the decided functionality that was agreed upon during those weeks--within 10 minutes.
That task would've taken me 3 days (3 days of 8 hours = 1,440 minutes)	Promptless AI created the first draft in 1% of the time it would've taken me. Then, I just needed to re-order the paragraphs in a logical way, clean up the formatting (tables, diagrams, etc), and match our company voice.

36% page reduction from IA

Atelio of FIS Global (Seattle) -- Revamped information architecture, reduced dev pages by 36%, made navigation easier

Before (May 2024)	After (July 2025)
PROBLEM: Short pages were hastily written by engineers who used more pages than needed Total of 50 pages	MY SOLUTION: I revamped and streamlined the information architecture to improve the flow and group related topics. Total of 32 pages

Revamped 60% of pages

Nium (San Francisco) -- Revamped 60% of the pages created by engineers hastily

Before (May 2024)	After (July 2025)
PROBLEM: Pages were hastily written by engineers just to have "something" documented about their products, such as: Payins Payouts Verify	SOLUTION: I revamped and organized the pages with many more details and related concepts as well as adding many new pages, such as: Payins          Foreign exchange        Use cases Payouts       Transactions                 Fees and limits Cards           Reports                         Nium portal Verify           Open banking               Developers

2x customers onboarded

Nium (San Francisco) -- Doubled the number of people onboarded per month with 75% fewer issues

Before (Oct 2022)	After (Dec 2022)
PROBLEM: New clients were unable to onboard themselves due to the unclear method to them–-and even to Nium. Each region (AU, EU, HK, SG, UK, US) contains five very complex spreadsheets describing various steps of onboarding for various client types and situations:	MY SOLUTION: I created a clear onboarding process with sections of customer types for common onboarding steps and for region-specific parameter and example pages. Immediately saw twice as many customers onboarded and only a quarter of the Helpdesk requests for onboarding

#1 cybersecurity coder

Yahoo (Sunnyvale) -- Among 2,733 Python developers, I was ranked as #1 in the Cybersecurity Code Warrior Challenge

Before (May 2020)	After (May 2020)
PROBLEM: 1. Find the 28 vulnerability types in 28 separate Python programs. 2. Then choose the best fix to block the vulnerability among 4 choices. 3. Incorrect guesses = fewer points. I had never taken a cybersecurity course but had read many articles related to it.	MY SOLUTION: - Since there was no time limit on the vulnerability questions, I had plenty of time to read and understand the code. - I was able to imagine what might cause a problem and then how to block it.
	I was #1 among 2,733 developers ... ... two weeks later, I was #42 among 7,624

40% fewer pages

Yahoo (Sunnyvale) -- Reduced obsolete pages by 40% and added automation to prevent obsolete pages

Before (May 2019)	After (Dec 2019)
PROBLEM: Too many obsolete and disorganized document pages. - Going to have an audit on all documentation pages. - Most in Confluence, needing to be in Markdown. - Many pages were obsolete but not clear which ones. - Many pages were not in easily discoverable places. - Many related pages/topics could be combined. - Moving forward, how to prevent "stale" pages?	MY SOLUTION: Implement reminders to review unmodified pages of a specified number of days. - I reviewed pages with SMEs - Archived obsolete pages - Merged similar pages - Organized pages by product - Migrated them to Markdown in Yahoo's GitHub - I suggested a system of tags on every page:     -Owner, LastModified, DaysTillStale - A daily script looks for pages that haven't been edited in that page's time limit and sends an email to the owner (or owner's manager) of that page to review it.
	- Total number of pages reduced by 40%. - Automated a timely reminder to page owners.

#1 dept in documention

Google (Mt. View) -- The AdWords API dept needed to improve the "freshness" of their internal documentation pages

Before (May 2018)	After (Nov 2018)
PROBLEM: 300+ Confluence documentation pages hadn't been updated in years, so many pages: - had outdated information - pointed to obsolete pages - or pointed to the wrong page	MY SOLUTION: I gathered the latest information, updated pages, combined similar pages, and archived obsolete pages. 100% of the AdWords API documentation pages were correct and were modified more recently than any other department within Google.

0 help from engineers

Couchbase (Santa Clara) -- Needed to install a pre-QA version to document it while no engineer was available to help

Before (Apr 2018)	After (Apr 2018)
PROBLEM: Need to document features before testing is done. - Couchbase Server v6.0 was still being coded. - Less than half of v6.0 had completed QA testing. - Documentation was needed for an event. - There wasn't an installed instance I could use.	MY SOLUTION: Install the Alpha version to use it and document it. (it was like installing Linux in 1996 before Google) 1. Find an unused server I could reformat. 2. Download and install Ubuntu 18.0. 3. Find, download, and install dependency files. 4. Find, download, and compile CB v6.0 source code. 5. Run Couchbase Server and create queries that use the new ANSI indexing and other new features.
	I was able to document pre-QA features in time for an event without help from the software developers.

80% time saved on writing

Couchbase (Santa Clara) -- Reduced the writing process of a new feature from 4-6 weeks to 4-5 days

Before (Mar 2017)	After (Aug 2017)
PROBLEM: Turnaround time to add features took way too long. 1. Writers authored a feature in Oxygen. (2-3d) 2. Webpage staged for engineer review. (2-3h) 3. Engineers eventually gave feedback. (3-5d) 4. Webpage updated. (1-2d) Time for iteration: 6-13d x 2-4 iterations (going to Step 2) ======================== Total of 4 - 6 weeks per feature	MY SOLUTION: Google Docs for synchronous writing/editing. 1. I authored a single feature in Google Docs and shared it with all engineers involved. (2-3d) 2. While engineers discussed how the feature will be finalized, I started on the next feature. 3. After the engineers finalized a feature's text, I imported it in Oxygen and staged it only once. (2-3h) ====================== Total of 4 - 5 days per feature
	Average time saved: 5 weeks → 1 week (80%)

30% fewer helpdesk tickets

Couchbase (Santa Clara) -- Reduced the number of helpdesk tickets by 30%

Before (May 2017)	After (Aug 2017)
PROBLEM: Customer Support had 200+ tickets to get through. The existing documentation: - didn't have examples - didn't explain in-depth enough	MY SOLUTION: I added: - many examples with Couchbase's sample database - more-detailed explanations to many concepts
	Helpdesk said the number of support tickets noticebly reduced by about 30% each week.

2 VPs with opposite priorities

Couchbase (Santa Clara) -- How to prioritize the writing of new features vs updating old pages? Google Analytics!

Before (Apr 2017)	After (May 2017)
PROBLEM: Two VPs had opposing priorities for me: - Engineering VP for new v4 features. - Product VP for improved v3 docs. - Every day they overrode the other.	MY SOLUTION: Google Analytics. It was clear: - About 85% of the documention pages viewed were v3. - Which pages were viewed the most often. - Which pages were looked at the longest time. So, I knew exactly how to prioritize which page to improve next.

50% less time + fewer errors

VA Medical Center (Boston) -- Reduced time to create clinical trials by 50% with 50% fewer errors

Before (Feb 2012)	After (Apr 2012)
PROBLEM: PMs complained that they needed to enter long, complex eDC expressions that were hard-to-read and very error prone.	MY SOLUTION: I suggested to the engineers to separate each element into dropdown boxes with human-readable values that automatically matched the level of parenthesis.
For example: NOT((intAge<="35") OR (ynSmoke !="1"))

1/3rd the time to edit eDC

VA Medical Center (Boston) -- Reduced the time needed to make any changes to an eDC clinical trial form to one-third.

Before (Aug 2011)	After (Dec 2011)
PROBLEM: There was no clear approval path when changing any part of a clinical trial's process or database or questionnaire. To add or modify a feature or request to their complex eDC system: - It went through a maze of approvals - Various people in 7 departments There was no clear path or order of who to talk with next when each step is approved or rejected.	MY SOLUTION: Over two months, - I talked with 20+ people - I refined a 7-department swimlane diagram - I made a 100-step complex-yet-clear flowchart that all department managers agreed on. The average time to make an edit to an eDC form went from 3 weeks to 1 week.

#1 in a Green campaign

HP (Singapore) -- 1st place in Green Environment Campaign for my ideas that save 20-30% of paper used worldwide

Before (Jan 2008)	After (Feb 2008)
PROBLEM: HP wants to reduce the amount of paper and ink used to be more "green". HP's Word DEFAULT.DOC file settings were: - 1.5” left, right, top, and bottom margins - 12pt font - Printers print singled-sided by default	MY SOLUTION: Slightly smaller margins and fonts. Change the DEFAULT.DOC file settings to: - 0.5" left, right, top, and bottom margins - 10pt font - Print double-sided by default
Example submitted: HP’s Orientation Table with URLs written out, printed on 14 pieces of paper	Became 2 pages (1 piece of paper) since: - 34% more print area (49.5 to 75 in2) - 17% more words per page (12pt to 10pt) - Saved 20-30% of paper printed
Cost to implement: 1 hour of administrator's time	Was 1 of 5 winners from 1,800+ submissions