Reporting

Unit economics

6 min read Reporting

What Kai actually costs you per outcome: per contact researched, per reply, per positive reply, per meeting, and once your CRM is mapped, per sales-accepted opportunity and as a percentage of the revenue it produced. It sits under Reporting > Economics and is visible to managers, admins, and superadmins only.

What the page answers

Most outbound tooling can tell you what it spent. This page divides that spend by the things you care about, so the number you quote in a budget conversation is a cost per meeting rather than a cost per seat.

Kai reporting, showing the Reporting group in the sidebar with Analytics, Economics, K.AI Analysis and Health.

Economics sits under Reporting, alongside Analytics, K.AI Analysis and Health. Analytics tells you what happened; Economics tells you what it cost.

The metrics

Total spend for the selected window.
Cost per lead, over contacts researched.
Cost per reply, over all replies received.
Cost per handraiser, over positive replies. This is usually the most honest early number, because it is the first point where the spend has produced something a human wants to act on.
Cost per meeting. The card says where the meeting count came from, either from your CRM or from Kai's own reply marking. Prefer the CRM figure once it is available; it is the one your finance team will accept.

Map your CRM properties and two more appear:

Cost per SAO, over sales-accepted opportunities.
Cost as a percentage of revenue across the contacts Kai sourced.

A metric with too little behind it says "not enough data" rather than showing a number. A cost per meeting computed from one meeting is not a unit economic, it is an anecdote with a currency symbol. Give the window enough volume before you quote anything from here.

Where the spend goes

The breakdown splits total spend three ways:

Generation. Writing the sequences.
Research. Understanding the account and the person before writing.
Firecrawl. Actually reading the web pages the research rests on.

There is also a per-model breakdown showing spend and call count for each AI model in use. That is the lever most people do not know they have: if one model is taking a disproportionate share of spend for a stage where a cheaper one would do, this is where you see it. Model choice is set in Connecting your stack.

Mapping your CRM outcomes

Cost per SAO and cost as a percentage of revenue cannot be computed from Kai's own data, because Kai does not know what your team accepted or what it closed. You supply that by pointing Kai at the HubSpot properties that carry it. Click Edit mapping and fill in:

Field	What it is for
Kai-sourced tag property	The property that marks a contact as having come from Kai. This is what scopes every CRM figure to Kai's pipeline rather than all of it.
Tag value that counts	Which value of that property counts. Leave blank to accept any value.
SAO property and SAO value	The property and value that mean sales accepted this opportunity.
Revenue property	A numeric property carrying deal value.
Meeting property and Meeting value	How your CRM records that a meeting happened. Once set, cost per meeting switches from Kai's count to your CRM's.

Save the mapping and the page reports what it found: how many SAOs and how much revenue across your Kai-tagged contacts. If the mapping is set but the pull fails, the page says so explicitly and shows the error rather than quietly displaying zero.

The tag property is the one that matters most. Get it wrong and you are dividing Kai's cost by your entire company's revenue, which produces a flattering number that falls apart the moment somebody checks it. Confirm the property and value with whoever owns your HubSpot before you present anything from this page.

Reading it honestly

Early numbers look bad, and that is expected. Research and generation spend land immediately; meetings land weeks later. A young workspace divides a full cost by a partial outcome. Judge it after a full sales cycle, not after a fortnight.
Compare against loaded cost, not licence cost. The comparison a buyer actually cares about is cost per meeting against what the same meeting costs through an agency, a contractor, or a rep's time.
Watch the ratio between stages. Research spend rising faster than generation spend usually means campaigns are sourcing far more contacts than get approved. That is a targeting fix, and it shows up here before it shows up anywhere else.
Use the window deliberately. A short window is volatile. Widen it before concluding a trend exists.

If the page is empty or wrong

Everything reads "not enough data". Not enough activity in the window. Widen it.
Cost per meeting looks impossible. Check which source the card names. Kai's own count only includes replies somebody actually marked as a booked meeting, so it undercounts when the team is not marking them.
SAO and revenue are missing. The mapping is not set. The page tells you so directly under CRM outcomes.
You cannot see the page at all. It is manager and above; a standard seat is redirected to Reporting. See Users & roles.