Call Tags

Call tags categorize the calls Glyphic processes. They drive scorecard selection, Custom Insight scoping, filtering on the call list and in Ask Glyphic, and reporting. Most calls are tagged automatically by AI based on the tag's name and description, but you can also tag any call manually.

Where to manage tags

Tags are configured at Settings → Call Tags. Each tag has:

• Name — what shows up on the call.
• Description — a short definition of when the tag applies. The AI uses this to decide whether to apply the tag, so be specific.
• Group (optional) — see below.
• Auto-tagging — when on, Glyphic predicts whether this tag applies to every call automatically.

Grouped vs. ungrouped tags

Tags can optionally belong to a group. The behavior is different in each case:

• Grouped tags are mutually exclusive. Exactly one tag from the group is applied to each call. If a different tag from the same group already exists, it's replaced. Use a group when the tags represent values of a single dimension — e.g. a Call Type group with tags like Discovery, Demo, Negotiation, or a Stage group with tags like S1, S2, S3.
• Ungrouped tags are independent. Any number can apply to a single call. Use ungrouped tags for orthogonal flags — e.g. Competitor mentioned, Pricing discussed, Champion identified.

Auto-tagging

When auto-tagging is enabled on a tag, Glyphic asks the AI to apply it (or, for grouped tags, to pick one tag from the group) after the call is processed. The AI sees the call transcript, your Product Description, and the tag names and descriptions. For richer context, it also looks at:

• Other calls on the same deal if the call is matched to a CRM deal.
• Other calls with the same company if the call is matched to a company but not a specific deal.
• Just the call itself if neither matches (cold calls, unknown participants).

It looks back up to 90 days and at most 30 prior calls. If the AI isn't confident about any tag in a group, it picks "Other" and no tag is applied — so it's normal for some calls to come back untagged on a particular dimension.

Why descriptions matter

Auto-tagging accuracy depends almost entirely on the tag's name and description. If you're seeing miscategorized calls, the fix is usually to make the description sharper rather than to add more tags. Describe what kind of call should and shouldn't get the tag, in your own language.

For example, instead of:

Negotiation — "Negotiation calls."

Write:

Negotiation — "A discussion between sales and the prospect after the demo, focused on price, terms, contract, or procurement. Excludes onboarding or post-sales conversations."

Default Call Type group

New organizations come with a Call Type tag group with auto-tagging enabled and these tags out of the box:

• Cold Call — cold outbound call.
• Discovery — initial discovery (qualification) call in a deal.
• Demo — call where the main agenda is demoing the product to the prospect.
• Deal Execution — sales discussions after the initial demo and before close (negotiation, finalization). Excludes post-sales activity.

If the default tags aren't matching your sales process — or you're seeing calls land in the wrong category — edit the names and descriptions to match how your team actually talks about call types, and any new calls will be retagged on that updated definition.

Manually overriding a tag

Anyone with access to a call can change its tags from the call page. For grouped tags, picking a different value from the group replaces the existing one. Manual changes take precedence over the AI prediction; we also keep a per-call audit trail of who changed what tags when.

If you change a call's tag to one that has a scorecard attached, the scorecard is recalculated automatically.

What tags are used for

• Filtering the call list. Narrow your call view to a specific tag (e.g. all Discovery calls in the last quarter).
• Scorecards. Each coaching scorecard is attached to one or more call tags. Only calls with a matching tag are scored against that scorecard, so you can have a different rubric for Discovery than for Demo, etc.
• Custom Insights. Each Custom Insight can be filtered to a set of tags, so the AI only tries to extract that field from relevant calls.
• Agents and Ask Glyphic. Tag filters are first-class arguments in agent prompts and Ask Glyphic queries (e.g. "Across my Discovery calls last month, what objections came up most?").
• Reporting. Tag-based call volume and trends in the Stats / Coaching views.
• Onboarding playlists. Use a tag like Great discovery as a curated training set for new reps.

Permissions

Tags are organization-wide. Anyone in the workspace can create, edit, and delete tags from Settings → Call Tags. Anyone who can see a call can change its tags.

Deleting a tag also removes it from any scorecards or Custom Insights that filtered on it, so calls that previously matched will simply stop being scored against those scorecards or having those insights extracted.