Back office ops · Production

Instacart's internal AI assistant Ava (GPT-4) is relied on by 43% of employees to save over an hour per week

The problem

Instacart employees across engineering, legal, marketing, and operations were spending significant time on tasks like assimilating unfamiliar codebases, resolving syntax errors, synthesizing meeting notes, and drafting communications, without a secure integrated internal AI tool.

Workflow diagram · grounded in source
1
Employee invokes Ava via Slack
trigger
“Ava's Slack plugin is invoked over 5,000 times in threads and group conversations”
2
AI-assisted code generation
ai_action
“60% of our engineers who generate around 70,000 lines of code using Ava every month”
3
Syntax error resolution
ai_action
“With Ava, we can input the problematic code and error messages, and in 80–90% of cases, it instantly provides a solution”
4
Internal knowledge search via Slack bots
integration
“Through custom-built Slack bots powered by Ava, we've optimized our ability to rapidly search internal resources such as Confluence/Wiki”
5
Meeting transcript to recap email
ai_action
“I can simply record the meeting with transcript generation enabled, upload the transcript to Ava, and request a concise meeting recap email. Ava delivers it complete with critical highlights, designated action items, next steps and ident…”
6
Human cross-checks Ava output
human_review
“I always ensure to cross-check any crucial information it provides with reliable sources, to maintain the accuracy and integrity of my work”
7
Prompt Exchange enables prompt sharing
feedback_loop
“a unique "Prompt Exchange" that encourages users to share and reuse prompts, enhancing collaborative productivity”
Reported outcome

Ava is relied on by 43% of Instacart's company to save over an hour per week, with 60% of engineers generating around 70,000 lines of code monthly using Ava.
Daily active users in the Brand Partnerships org surged 50% and daily interactions grew approximately 40% after an internal best-practices roadshow.

Reported metrics
company relying on Ava weekly43%
Time saved per user per weekover an hour a week
engineers using Ava for code generation60%
Lines of code generated monthly70,000
Show all 12 reported metrics
company relying on Ava weekly43%
time saved per user per weekover an hour a week
engineers using Ava for code generation60%
lines of code generated monthly70,000
Slack plugin invocations per month5,000
threads and channels summarized per month200
syntax error instant resolution rate80–90%
Brand Partnerships daily active users increase50%
Brand Partnerships daily interactions increaseapproximately 40%
team efficiency and productivity increase20–30%
job application volume increasetriple-digit increase
email writing prompt usage count150
Reported stack
AvaGPT-4SlackConfluenceWiki
Source
https://tech.instacart.com/unlocking-efficiency-how-ava-became-our-ai-productivity-partner-f1a560686361
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Ava is relied on by 43% of Instacart's company to save over an hour per week, with 60% of engineers generating around 70,000 lines of code monthly using Ava.

What tools did this team use?

Ava, GPT-4, Slack, Confluence, Wiki.

What results were reported?

company relying on Ava weekly: 43%; Time saved per user per week: over an hour a week; engineers using Ava for code generation: 60%; Lines of code generated monthly: 70,000 (source-reported, not independently verified).

How is this back office ops AI workflow structured?

Employee invokes Ava via Slack → AI-assisted code generation → Syntax error resolution → Internal knowledge search via Slack bots → Meeting transcript to recap email → Human cross-checks Ava output → Prompt Exchange enables prompt sharing.

Related back office ops cases