Why should you care?

CRM copilot prototypes and coding assistants often start on OpenAI before enterprise moves deployments to Azure OpenAI.

See it live — copy this example

Paste into an ASP.NET Core 8+ / AgentNest project, then run with dotnet run (set your API keys in user-secrets).

// AgentNest.Api/Services/OpenAiLeadTagger.cs
public sealed class OpenAiLeadTagger(IChatClient chat)
{
    public async Task<IReadOnlyList<string>> TagAsync(string notes, CancellationToken ct)
    {
        var msg = $"From CRM notes, return 3 tags as comma-separated words only:\n{notes}";
        var reply = await chat.GetResponseAsync([msg], cancellationToken: ct);
        return reply.Text.Split(',', StringSplitOptions.TrimEntries | StringSplitOptions.RemoveEmptyEntries);
    }
}

What happened?

• OpenAiLeadTagger uses the injected OpenAI-backed IChatClient.
• Parsing comma-separated tags keeps CRM storage simple without JSON schema overhead.

Practice next

1. Register IChatClient with AddOpenAIChatClient in Program.cs.
2. build OpenAiLeadTagger and register as scoped.
3. Map POST /api/crm/leads/tag with LeadNotesRequest body.
4. Ask for JSON array output and deserialize with System.Text.Json.
5. Add a max note length validator at 2,000 characters.

Remember

OpenAI integrates through IChatClient, not ad-hoc HTTP in controllers. Keep prompts small and parse predictable formats. Use OpenAI for dev; plan Azure for regulated tenants.

CRM auto-tagging pilot

Marketing wants automatic topic tags on inbound lead notes.

Outcome: OpenAiLeadTagger ships in staging within days using a shared API key vault secret.