Legal document review · Production

Build vs. buy AI agents: What 1,000+ enterprise deployments taught us about the real costs

The problem

Companies deploying AI agents consistently underestimate the true cost of building in-house, facing hidden burdens in time-to-value, compounding maintenance, and opportunity cost diverted from core business.

First attempt

Doctolib's internal ChatGPT-based tool DoctoGPT rapidly attracted 800 active users but became unsustainable as feature requests overwhelmed the team. Wakam's data science team built a working RAG chatbot prototype but found that keeping pace with the AI market would require tripling team size.

Workflow diagram · grounded in source
1
Leadership faces build-or-buy decision
trigger
“every leadership team eventually faces the same question: should we build this ourselves or buy a platform?”
2
Internal AI prototype built
ai_action
“their five-person data science team began building a custom AI chatbot with RAG capabilities. While they had the expertise to create a working prototype, maintaining it required constant engineering effort”
3
Feature demand overwhelms team
feedback_loop
“We created a Feature Requests JIRA board that revealed massive demand. We were overwhelmed with requests within days of launch”
4
Company adopts Dust platform
integration
“Wakam recognized these limitations early and pivoted to Dust. Within two months they achieved 70% employee adoption and deployed 136 AI agents across the organization”
5
Business outcomes delivered
output
“Their legal team cut contract analysis time by 50%, and data teams enabled self-service intelligence that dramatically reduced processing time”
Reported outcome

Companies that adopted Dust achieved rapid deployment and broad employee adoption: Wakam hit 70% adoption in two months with 136 agents deployed and 50% reduction in legal contract analysis time; Ardabelle evaluated 50% more deals in the same timeframe; CMI Strategies achieved 95% adoption with 60-70% time savings on commercial proposals.

Reported metrics
Wakam employee adoption70%
Wakam AI agents deployed136
Wakam contract analysis time reduction50%
Ardabelle analyst queries per week150+
Show all 17 reported metrics
Wakam employee adoption70%
Wakam AI agents deployed136
Wakam contract analysis time reduction50%
Ardabelle analyst queries per week150+
Ardabelle deals evaluated increase50%
Ardabelle adoption rate95%
Ardabelle adoption timeline90 days
CMI Strategies consultant adoption95%
CMI Strategies commercial proposal time savings60-70%
CMI Strategies executive summary speed improvement50%
CMI Strategies monthly cost per consultant30-40 euros
CMI Strategies ROI timelinemeasurable ROI within weeks
November Five agent deployment time20 minutes
DoctoGPT active users800
Build timeline underestimation6-12 months
Dust total enterprise deployments1,000+
Mirakl employee transformation goal75%
Reported stack
DustGPT-4ChatGPTRAGZendeskSlackNotionGitHubSalesforceJIRAvector databases
Source
https://blog.dust.tt/build-vs-buy-ai-agents/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

Companies that adopted Dust achieved rapid deployment and broad employee adoption: Wakam hit 70% adoption in two months with 136 agents deployed and 50% reduction in legal contract analysis time; Ardabelle evaluated 5…

What tools did this team use?

Dust, GPT-4, ChatGPT, RAG, Zendesk, Slack, Notion, GitHub, Salesforce, JIRA.

What results were reported?

Wakam employee adoption: 70%; Wakam AI agents deployed: 136; Wakam contract analysis time reduction: 50%; Ardabelle analyst queries per week: 150+ (source-reported, not independently verified).

What failed first in this deployment?

Doctolib's internal ChatGPT-based tool DoctoGPT rapidly attracted 800 active users but became unsustainable as feature requests overwhelmed the team.

How is this legal document review AI workflow structured?

Leadership faces build-or-buy decision → Internal AI prototype built → Feature demand overwhelms team → Company adopts Dust platform → Business outcomes delivered.

Related legal document review cases