HelpSpot

@HelpSpot search for all open requests from acme.com in the last 30 days and group them by category

Found 14 open requests from acme.com customers since January 15. Grouped by category: Billing (6 requests) — mostly questions about invoice timing and payment methods; Technical Support (5 requests) — API authentication errors and webhook delivery issues; Account Management (3 requests) — seat additions and plan changes. The oldest unresolved request is #8472 from January 18, still assigned to the billing queue with no staff response yet.

This example shows the MCP's search tool paired with AI summarization. The search returns raw request metadata; the AI groups and interprets it. Note that HelpSpot's search is text-based — complex date-range or multi-field queries may require multiple calls or filter-based approaches instead.

@HelpSpot create a new request for customer sarah@techcorp.io with subject 'API rate limit clarification' and add an internal note that this is a pre-sales question

Created request #9284 for sarah@techcorp.io with subject 'API rate limit clarification'. Added internal staff note flagging this as pre-sales. Request is now in the default inbox queue. HelpSpot's configured automations may have triggered — check your notification settings if you expect an auto-assignment or customer confirmation email.

Demonstrates the create-request tool, which writes directly to your HelpSpot instance. The MCP does not control what happens after creation — automations, workflows, and notification rules in your HelpSpot configuration determine whether emails are sent or staff are assigned. Always verify side-effects in a test category first.

@HelpSpot pull my top 10 saved responses and draft a reply to request #9201 using whichever template fits best

Retrieved your 10 most-used saved responses. Request #9201 is a billing question about invoice delivery timing. Best match: template 'Billing – Invoice Schedule Explanation' (you've used this 47 times). Draft reply: 'Hi [customer name], invoices are generated on the first business day of each month and emailed to your billing contact within 24 hours. If you need a copy sooner, I can send your current invoice now — just let me know.' Adjust placeholders before sending.

Shows the list-saved-responses tool combined with request retrieval and AI reasoning to pick the right template. The MCP returns template text with placeholders intact; you must review and personalize before posting the reply. HelpSpot does not auto-send — this is a drafting workflow, not a one-click response.

When HelpSpot wins for 24/7 support team transitions

A 6-person support team running three 8-hour shifts needs clean handoffs without Slack chaos. The HelpSpot MCP pulls the inbox filter at shift start, surfaces requests that aged past SLA thresholds, and lets the incoming agent scan saved responses before replying—all inside the AI workspace where the team already documents escalation notes. This works when your ticket volume sits under 200 active requests and your staff actually uses HelpSpot's saved-response library; above that scale the search tool gets slow and you're better off keeping HelpSpot open in a browser tab. If your handoff ritual is already working and takes under 10 minutes, skip the integration—Switchy shines when the current process involves toggling between HelpSpot, Slack, and a wiki.

Pull request history for account health scoring

A 3-person customer success team preparing quarterly business reviews needs to score account health by support load, and their CRM doesn't sync HelpSpot request counts. The search and get-request tools let the team pull a customer's full ticket history, count escalations, and draft the health narrative in one Switchy thread without exporting CSVs or writing SQL. This scenario works when you have under 50 accounts to review and your HelpSpot categories map cleanly to health signals like 'bug', 'feature request', or 'onboarding question'. If you're reviewing hundreds of accounts or need pivot tables, the MCP's cursor pagination becomes tedious—export to a spreadsheet instead. The buying call: if you're copy-pasting ticket URLs into Google Docs today, the MCP saves 15 minutes per review.

When HelpSpot helps ramp a junior agent faster

A support manager onboarding a new hire wants the agent to learn the team's response patterns before taking live tickets. The list-saved-responses and list-filters tools surface the ten most-used templates and the team's queue structure in a single Switchy prompt, so the manager can walk through triage logic without screen-sharing HelpSpot for an hour. This works when your HelpSpot instance has under 30 active filters and your saved responses follow a naming convention the new hire can parse; if your filter list is a junk drawer of 80 abandoned experiments, the MCP output overwhelms instead of teaching. The trade-off: this replaces a live walkthrough, not your actual training docs—use it to accelerate day-two shadowing, not day-one compliance reading.