Does the company need Google Workspace for Gemini Enterprise?

No. Gemini Enterprise is connected to Microsoft 365 and the CRM via ready-made connectors and to SAP via a custom connector; the existing systems stay.

Does the AI decide on quotes?

No. The agents prepare: they review tenders and create drafts. Every quote is checked and approved by sales.

Why two tools instead of one?

The tender agent can be built with the toolkit in Gemini Enterprise. The quote agent needs configuration rules and an SAP connection, so it is developed on the Gemini Enterprise Agent Platform. Sales uses both in the same interface.

Which AI models are used?

Gemini by default. For analysing very long tender documents, a Claude model from Anthropic can also be selected in Model Garden.

How long does a project like this take?

The project plan allows around four months: two weeks each for concept and analysis, four weeks for the Gemini Enterprise pilot, four weeks for the tender agent and six weeks for the quote agent. The actual duration depends on scope, the number of use cases and the system landscape, and may differ; we set the exact schedule in the concept.

Mechanical engineering · AI and agents

Gemini Enterprise in sales: reviewing tenders and preparing quotes in mechanical engineering

A machine builder with around 1,200 employees, about 100 of them in sales, introduces Gemini Enterprise for the sales team. One agent reviews incoming tenders and flags requirements, risks and open questions; a custom agent on the Gemini Enterprise Agent Platform creates a first draft quote from configuration and SAP data. The project plan allows around four months.

Sales organisations with many tenders and technical quotes
Companies with Microsoft 365 and SAP that want to use AI without switching platforms
Management teams that want AI with measurable benefits instead of isolated pilots

Arrange an initial consultation

Why does a machine builder use AI in sales?

Every week, the sales team responds to several tenders and prepares technical quotes. Much of the time goes into reading extensive specifications, checking them against the company's own product range and gathering prices and delivery times from SAP. Management wants to win back this time. What matters is not another AI pilot but noticeable relief in day-to-day business. At the same time, Microsoft 365 and SAP are to stay as they are.

Knowledge is scattered.

Previous quotes sit in SharePoint, customer data in the CRM, prices and configurations in SAP. Anyone preparing a quote searches several systems at once.

Approvals stay with people.

Every quote has to be checked technically and commercially before it goes out. AI is meant to prepare, not decide.

Project profile

1,200

Employees

around 100 people

Sales

Mechanical and plant engineering

Industry

Microsoft 365, SAP, CRM

Systems

several per week

Tenders

Technologies used

Gemini EnterpriseGemini Enterprise Agent PlatformAgent DesignerAgent Development KitGemini and Claude

Approach

How do Gemini Enterprise and GEAP work together?

Gemini Enterprise is the workspace for the entire sales team. It searches SharePoint and the CRM with the existing permissions, and SAP data is added via a custom connector. This lets it answer questions about previous quotes, customers and products. A tender agent, built without code in Agent Designer, reviews new specifications, checks them against the product range and flags requirements, deviations and open questions.

A toolkit isn't enough for preparing quotes: this needs configuration rules, prices and delivery times from SAP. That's why a custom quote agent is built on the Gemini Enterprise Agent Platform. It creates a first draft quote, which sales checks and approves. For analysing very long documents, another model can also be selected in Model Garden, such as Claude from Anthropic.

Arrange an initial consultation

Concept and architecture

Target picture for the AI platform: identities and permissions, access to SharePoint, CRM and SAP, governance rules, model choice and cost framework. Two weeks.

Use cases and data

Capture sales processes, measure the tender and quote workflow, clarify data sources and permissions. Two weeks.

Gemini Enterprise for sales

Connect SharePoint and CRM via ready-made connectors and SAP via a custom connector, pilot with 20 sales staff, training. Four weeks.

Tender agent

Built in Agent Designer, tested with previous tenders, then released to the entire sales team. Four weeks.

Quote agent on GEAP

Developed with the Agent Development Kit, connected to SAP, safeguarded with test cases and an approval step. Six weeks.

Operations and expansion

Monitoring, cost control and further agents, for example to prepare customer meetings.

Which technology is used?