SharePoint CRM built with Copilot Code for about $55

Key insights

• Anderson used Copilot Code plan mode to propose the CRM structure before building.
• The proposed design included 13 SharePoint lists and three document libraries.
• The CRM covers contacts, companies, deals, activities, tasks, documents and playbooks.
• Copilot Code requested approval before creating lists and using Microsoft Graph.
• The /cost command reported 5,415 credits, which Anderson described as roughly $55.
• The app was still in preview and had not been finished or published in this video.

Daniel Anderson [MVP] recently published a first look at using Copilot Code to build a HubSpot-style customer relationship management system with SharePoint as its back end. The demonstration moves from planning the information architecture to generating an application, with Copilot Code reporting a build cost of 5,415 credits, roughly $55. The main takeaway is that the video shows an early, unfinished build—not a published CRM ready for general use.

What's new

Copilot Code appears in Anderson’s demonstration alongside Chat and Cowork in Microsoft 365 Copilot. He selects a model and begins in plan mode, asking it to prepare the structure for a CRM before creating the solution. The resulting plan proposes 13 lists, three document libraries, relationships between records and an implementation sequence.

The proposed SharePoint structure includes lists for contacts, companies, deals, activities and tasks. Document libraries are intended for files and playbooks. Anderson reviews the plan, approves it, and moves Outlook integration to a later phase rather than including it in the initial build.

After that approval, Copilot Code begins creating the solution. The video shows a first look at companies, deals, record pages and documents. It also shows the tool asking before it creates lists and before it uses Microsoft Graph, a Microsoft service for working with data across Microsoft products.

Why it matters

For SharePoint makers and business teams, the demonstration is relevant because it shows an AI-assisted route from a description of a business need to a proposed data structure and a preview application. The plan-first step gives the user a chance to review what Copilot intends to build before it starts creating the underlying SharePoint components.

That can help people assess whether a proposed application matches their process before committing to the build. In this example, separating Outlook integration into phase two is a concrete scope decision: the initial CRM is built around SharePoint lists and libraries, while that integration is deferred.

The approvals also matter for administrators and users responsible for the SharePoint environment. The demonstration shows Copilot Code pausing for approval before creating lists and using Microsoft Graph. It does not explain the permission model, administrative controls or how approval behavior can be configured, so the video should not be treated as a complete governance walkthrough.

The cost figure needs similar context. Anderson’s run used 5,415 credits, which he estimates at roughly $55. That is the reported cost of this particular build, not a quoted standard price or a promise that other projects will cost the same.

How it works

1. Anderson chooses Copilot Code and starts in plan mode, asking for the CRM structure before the application is built.
2. Copilot Code returns a proposed architecture: 13 lists, three document libraries, record relationships and a suggested order for implementation.
3. He reviews and approves the proposal, while moving Outlook integration to phase two.
4. Copilot Code proceeds with the build and asks for approval before creating lists and using Microsoft Graph.
5. Anderson inspects the preview, including company and deal views, record pages and documents, then uses the /cost command to see the credit usage.

This sequence shows a human reviewing the plan and responding to approval requests rather than simply accepting an end-to-end build without checkpoints. The video presents the generated application as a preview; it does not show the later changes and publication process.

Limitations and open questions

Anderson explicitly describes this as Part 1. The app works in preview, but it is not finished or published in the video. He says a later part will cover a fuller review, changes and publication; those steps are outside this demonstration.

The material does not establish what licensing is required, whether Copilot Code is available to a particular tenant, or what controls an administrator needs to enable it. It also does not provide enough detail to evaluate the CRM’s security, data quality, performance, long-term maintenance or suitability as a replacement for a dedicated CRM product.

The roughly $55 estimate is specific to the displayed run and its credit usage. The video does not establish a standard price for building a CRM, nor does it verify how costs would change with different requirements. Treat the result as a prototype demonstration and assess the finished solution, permissions and operating costs before relying on it for business records.

Sources

• Daniel Anderson’s Copilot Code SharePoint CRM video

Frequently asked questions

What did the CRM include?

The proposed SharePoint structure included lists for contacts, companies, deals, activities and tasks, plus document libraries for files and playbooks. The video also previews company and deal views, record pages and documents.

Does the video show a finished, published CRM?

No. Anderson says the app works in preview but is not finished or published in this first part.

Is $55 the standard cost to build a CRM with Copilot Code?

The video reports 5,415 credits for this build, estimated at roughly $55. It does not establish a standard price for other projects.

What licensing or tenant setup is required?

The video does not specify licensing requirements, tenant availability or the administrative setup needed.