Client site tech stack console for agency account managers

Build me an internal console for my web agency where our account managers can see the technology behind every client site we manage. Today they re-run these lookups by hand the night before each quarterly review, and I want that ritual replaced by a screen that is always ready. There is no schedule and no trigger: a person opens this app and works in it.

Roster view (the landing page). Read our client roster from a Google Sheet using Get Values. The sheet has one row per client with the client name, their website address, the Slack channel we use with them, the retainer tier, and the account manager who owns the relationship. Let me set the spreadsheet id and range in settings and remember them. Render the roster as a table with one row per client showing client name, domain, retainer tier, owning account manager, the count of technologies detected, and the gap badges described below. Let the account manager filter the roster to just their own clients and search by client name or domain.

Stack refresh. A Refresh stack button on the roster view expands every client row into that domain's live technology profile using the BuiltWith Domain API Lookup. Important: the Domain API accepts up to 16 comma-separated domains in a single request and charges one credit per domain returned, so batch the roster into groups of 16 rather than firing one call per client. Store the returned profile per client so reopening the console does not re-spend credits, and show when the roster was last refreshed. Include a NOPII option in settings that strips personally identifiable information from the response, for agencies that would rather not pull personal data into an internal tool.

Gap badges. On each roster row, show coloured badges for the gaps that matter to our retainers, computed from the detected technology categories: no analytics, no tag manager or consent tool, an ageing content management system, no email or marketing automation, and no conversion tracking on a site that is running ecommerce. Use a clear colour scale so a red badge reads as a real gap and a neutral badge reads as informational, and let me hover a badge to see exactly which detection (or absence of one) triggered it. Ageing CMS should be judged from how long the CMS has been in place and whether the detected version or product is behind, so explain the reasoning in the tooltip rather than just asserting it.

Client detail view. Clicking a roster row opens that client's page with the full current stack grouped by category (analytics, CMS, ecommerce, advertising, hosting, javascript libraries and so on), each entry showing when it was first detected and when it was last detected. The Domain API returns these timestamps as epoch milliseconds (FirstIndexed, LastIndexed, FirstDetected, LastDetected), so format them as readable dates and add a relative hint such as added 3 months ago. Sort so the most recently added technologies float to the top, because a tool that appeared since our last review is exactly the signal an account manager needs: it usually means another agency sold the client something. Also show the company detail from the Domain API Meta block (company name, address, phone, social profiles) alongside the stack, respecting the NOPII setting.

Recommendations panel. On the client detail page, show technologies this site should consider adopting using the BuiltWith Recommendations API Lookup. The Recommendations API accepts root domains only, so normalise the sheet value before lookup: strip any protocol, path, and subdomain so shop.clientname.com is looked up as clientname.com. For each recommendation, pull adoption context from the BuiltWith Trends API so the account manager can see how widely adopted the technology actually is across the index and tell a real pitch from a fringe tool. The Trends API is free of credit charge, so this panel can refresh freely and does not need caching for cost reasons. Show each recommendation with its category, the adoption figure, and a simple read such as widely adopted or niche.

Pitch shortlist to Slack. Let the account manager tick the recommendations worth pitching for that client, optionally add a sentence of their own framing, and preview the message. Sending posts the shortlist into that client's Slack channel using the Slack Bot Send a Message action, with the channel taken from the client's row in the sheet and overridable in the UI. Format the message so it opens with the client name, lists each ticked recommendation with its adoption context and the gap it addresses, and reads as evidence rather than opinion. Nothing is ever posted automatically: sending is always a deliberate click. Record which recommendations were pitched and when, and show that history on the client detail page so the next quarterly review does not repeat a pitch the client already declined.

Persistence and per-user behaviour. Save the last fetched stack profile per client, the pitch history, any notes an account manager writes on a client, and the roster filter each user last used. The sheet stays the source of truth for which clients exist, so a client added to the sheet appears on the next roster load without any setup in the app.