Uber’s ad sellers will be 83% faster with agentic RFP intake and processing

Ad sales slowed by manual work

Uber is the largest mobility and delivery platform in the world, with 189 million active consumers each month and ads enabled in more than 40 countries. This global reach — combined with real-time data on rider and eater behavior — makes it a powerful channel for brands looking to reach customers at the moment of decision.

That demand translated into a steady stream of RFPs from prospective clients. To respond, Uber’s ad sellers often had to dig through documents that averaged more than 20 pages in various formats to find the details they needed.

After campaigns ended, sellers had to track down those same RFPs again to create wrap reports. The hours spent searching for key information could have been spent building client relationships and driving revenue.

From RFP to opportunity in minutes

To speed up how campaigns are sold and executed, Uber built agentic RFP processing with Agentforce that turns long, complex proposals into ready-to-act opportunities. Instead of manually reading 20+ page documents and entering data by hand, sellers can upload an RFP and let Agentforce do the heavy lifting.

When a seller uploads an RFP file to Salesforce, Agentforce will extract key details like campaign dates, budget, target audience, and product needs. It will instantly generate a clear summary, create a new opportunity in Agentforce Sales, and build a proposed media plan in Agentforce Media — complete with quotes and line items.

If sellers have questions or want to make changes like refining a summary or pulling specific insights, they can start a chat with Agentforce for help without ever leaving the sales CRM.

Agentforce will go a step further by acting as a strategic advisor. Based on the account, sales team, and RFP content, it will recommend the right advertising products and can update fields like close date and deal size. It will also make it easy to generate wrap reports at the end of campaigns by summarizing past RFPs — no more digging through old files.

From RFPs to action

Agentforce will work by turning unstructured RFP documents into structured data inside Salesforce.

When a seller uploads an RFP — whether it’s a 25-page PDF, slide deck, or email — to an account record, an external API will extract the text using a managed AWS Lambda function that connects with Google Cloud. That text will then be written back into a dedicated “RFP Text” field, where Agentforce can access and analyze it.

From there, a seller can start a chat within Agentforce Sales and ask, “Can you summarize this RFP?” Or they can type, “Create an opportunity,” and Agentforce will analyze the extracted text to identify key details like campaign dates, budget, audience, and goals. It will combine this with account context, like the customer’s assigned sales pod, and will apply a decision tree to recommend the right products.

Strong guardrails ensure Agentforce's accuracy. It bases replies only on extracted RFP data and Salesforce fields and follows strict logic based on specific sales pods, so every recommendation is relevant and realistic. A future enhancement will add a real-time inventory check before products are recommended.

Once the seller approves the plan, Agentforce will create an opportunity in Agentforce Sales, populate key fields like close date and amount, and build a proposed media plan in Agentforce Media with recommended products and line items. It will also generate a concise summary sellers can use immediately — or later for wrap reports.

Faster intake, faster deals

What once took sellers around 30 minutes — reading, extracting, and entering RFP details — will take an estimated five minutes, an 83% reduction in time spent on intake and analysis.

That efficiency matters because RFPs contribute to a significant share of Uber’s ad sales pipeline. In the U.S. and Canada, about 42% of opportunities in strategic verticals originate from RFPs, while globally, over 25% of consumer packaged goods deals start the same way. By accelerating how quickly these documents are processed, sellers can respond faster, build pipeline sooner, and move deals forward more efficiently.

Early results show that by removing a key bottleneck, Agentforce will help turn more RFPs into active opportunities — without adding extra work for sellers.

Salesforce Agentic Enterprise architecture

“The beauty of Agentforce is how deeply it’s integrated into the Salesforce ecosystem,” said Saurabh Sharma, Technical Lead Manager. With apps, data, and AI all on one unified platform, Uber’s staff can work faster without having to switch between tools and systems.

Implementation partnership

Uber partnered closely with Accenture and Salesforce Professional Services to build their RFP agent. Accenture co-built the agent and handled complex technical architecture items such as the AWS Lambda-to-Google Cloud conversion needed for text extraction.