See what your company
is doing with AI.Depthdata
The AI line item, finally defensible.
Depthdata reads the consoles your vendors already give you and returns a single figure a finance team will sign: what went out, what came back, and how much of that was counted rather than guessed.
Modelled Not a customer. Depthdata has no customers yet. Every figure below is the one the product demo ships, carrying the label the app would give it.
Targets · Q33 off target
Weekly adoption to 85%BEHIND
Cost / user to $200ON TRACK
Stay under $60k a monthAT RISK
Activate all Sales seatsAT RISK
Adoption and spend both trace to one cause.
The problem / 02
Four consoles. Four definitions of “active”. No answer.
A company buys ChatGPT, Claude, Copilot and Gemini. Four admin consoles, four billing pages, four ideas of what a user doing something means. The CFO asks one question, and answering it takes a week of exports.
“What did we get for the $48.2k?”
Nobody could say without a week of exports. By the time they could, the figure had moved.
\[ STAT NEEDED \]
Annual enterprise spend across AI tools
Needs a figure you would cite in a sales call. An analyst report you can link, or your own survey. Do not publish this section until it is filled.
\[ STAT NEEDED \]
Share of seats that go unused
Same test. If no credible public figure exists, cut the claim and lead with the workspace below, which needs no source.
\[ STAT NEEDED \]
Time to produce an AI usage report today
The cost of the problem in hours. A number from your own pipeline conversations is stronger here than an analyst's.
The tools report usage. Nobody reports whether it was worth it.
Current solutions / 03
Every vendor gives you a console. None of them gives you an answer.
Each one is good at exactly one tool, and each one defines its own terms. Put four side by side and you get four numbers that cannot be added up.
What each console covers, and what it cannot see
Measured
| Where you look today | What it covers | What it cannot tell you | | -------------------- | ------------------------------------------------------------------ | ----------------------------------------------------------------------------------------- | | ChatGPT admin | Seats, spend and audit events, through the official API | Usage only through exports and logs. Nothing about your other five tools. | | Claude console | Seats, usage, spend and audit, all through the API | Its own definition of an active user. Not comparable to the console next to it. | | Lovable | Seats through SCIM 2.0, on Enterprise with SSO | No usage or spend API. Credits live in the app, and audit is UI-only with no export. | | Gemini admin | Workspace seats and usage | Spend lives in Cloud Billing, a different system with a different owner. | | Vercel | Team members and roles by API, spend through the FOCUS billing API | Usage needs AI Gateway tagging on Pro and up. Audit only streams to a SIEM, with no pull. |
Source: the connector matrix Depthdata publishes, which states what every vendor exposes including where they expose nothing yet.
The opportunity / 04
Nobody needs a fifth dashboard. They need one definition.
Every AI vendor measures something different, and every one of them is measuring their own product's success, not yours. The gap is not more charts. It is a single definition of active, spend and value that holds across all of them, and that a finance team will sign off on.
The timing is not a hunch. It is in the matrix: vendors have started shipping read-only admin APIs, and a normalizing layer is only possible once they do. Where they still expose nothing, we say so.
Why now
Admin APIs exist. Six of eight tracked vendors expose seats, usage and spend today.
Budgets moved from experiment to line item, so someone now has to defend them.
Two of eight expose nothing but seats. That gap is published, not hidden.
The solution / 05
One schema. One baseline. Every number labelled.
It attaches to each vendor console with read-only scopes, translates their units into a single set of terms, and grades progress against what this workspace was doing in its first month. Nobody else's average is involved, so nobody can dismiss the result as someone else's company.
Measured A vendor's own API said so. Seat counts, timestamps, invoices.
Derived Arithmetic on the above. The formula is published, so you can redo it.
Modelled An estimate. The method is written down and the range travels with it. Never dressed up as a count.
Other tools show you a number. This one shows you how much to trust it.
That is the whole product in one sentence. A dashboard that cannot tell you which of its figures were counted and which were estimated is asking you to take all of them on faith, and no CFO signs that.
Product walkthrough / 06
From four consoles to one defensible number.
Five steps. Each one answers a question the last one raised, and each one is published, so a sceptical analyst can follow the arithmetic end to end.
01 / ATTACH
Scopes, not software
A read-only grant per vendor. Nothing lands on anybody's laptop, so nothing needs approving.
02 / TRANSLATE
Six dialects, one language
Each vendor counts in its own units. They get rewritten into terms that can be added together.
03 / ANCHOR
Graded against you
Progress is scored against what this workspace did in month one, team by team.
04 / BOUND
A range, never a point
Anything estimated arrives with the width of its uncertainty attached.
05 / GRADE
Arithmetic you can audit
The health figure is a formula, and the formula travels inside every export.
Seats, sessions, invoices. Nothing anybody typed.
What a person wrote into a chat box is not an input to any figure on this page, and cannot be. Where a vendor's feed carries that text anyway, it is dropped before it is written down. There is no setting to turn that off, which is the point.
Key features / 07
Four questions. Four instruments.
01
Executive dashboard
“Can I put this in front of the board?”
Four figures, each labelled. A board member reads it in thirty seconds and knows which numbers were counted and which were estimated, without asking anyone.
Business value The AI line item becomes defensible in a budget review instead of a fight.
Active weekly
0/ 287 seats
Measured
Straight off the APIs. Up 12 points in thirty days.
Workspace health
0/ 100
Derived
A formula over three inputs. It travels inside the export.
Time saved
0hrs/mo
Modelled
±140 hr. We will not collapse that to one figure.
Return on AI
0.0×
Modelled
$489k estimated back, on $144k out, this quarter.
02
Adoption by team
“Who is actually using this?”
A company average is an average of a healthy population and a stalled one. Split by department, six teams sit at or above the line. Two do not, and the average was hiding both.
Business value Enablement goes where it changes the number, not where it is easiest to run.
Active seats by department
Measured
Engineering68 / 74
Support40 / 45
Product28 / 32
Design20 / 24
Marketing21 / 26
Operations12 / 22
Finance15 / 28
Sales14 / 36
Company218 / 287
GRADED AGAINST / what each of these teams was doing in month one. No sector average is involved, so none of this is arguable.
03
Waste, priced line by line
“What can we cut without breaking anything?”
Every item traces to named seats and their activity, not to a percentage someone assumed. You can hand the list to procurement and defend each row.
Business value $4.04k a month recoverable here, with zero power seats touched.
Ranked by recoverable $ per month
Derived
Enterprise tier at free tier intensity
64 seats · under 12 prompts a week
$0.00k
Dormant ChatGPT seats
13 seats · no prompts in 30 days
$0
Idle Copilot licenses
18 seats · no AI assisted commits in 60 days
$0
Duplicate coverage
9 people · Cursor and Copilot
$0
What we did NOT recommend
Modelled
A full downgrade saves 31%, roughly four times the number above. It also breaks 109 power seats whose premium usage is justified by workload. The tool reports the bigger number and then argues against it.
UPPER BOUND 31% · BREAKS 109 POWER SEATS · WE WILL NOT ROUND UP
04
Goals and root cause
“Why are we going to miss?”
Two goals were failing for what looked like unrelated reasons: adoption was short, spend was climbing. Both traced to the same 22 seats. Two dashboards would have shown two problems.
Business value One move fixes both, instead of two teams working the same cause twice.
Weekly adoption to 85%AT RISK
76% now. Needs 9 points.
Stay under $60k a monthAT RISK
$48.2k now, trending up 14%.
One root cause
0 of 36
Sales seats sit dormant
They pull the adoption average below target and waste committed spend. Activating Sales is the single move that shifts both goals.
Product decisions / 08
What we left out is the product.
Every one of these cost us a feature a competitor lists on their pricing page. Each was worth it, and here is what it bought.
Decision
Label every number's confidence
Measured, derived or modelled, on the figure itself rather than in a footnote.
Trade-off Our headline numbers look smaller than a competitor's, because theirs do not separate the counted from the estimated.
Outcome Finance signs it. That is the only audience that matters for this line item.
Decision
Five executive figures, not fifty
The dashboard leads with four numbers and a health score. Everything else is one level down.
Trade-off Power users have to click. We lose the demo moment where a screen looks impressively dense.
Outcome The screen survives a board meeting, where nobody scrolls.
Decision
Report the smaller saving
A full downgrade saves 31%, four times what we recommend. We report it and then argue against it.
Trade-off The pitch is weaker. 31% is a better headline than 8.4%.
Outcome The 8.4% is the part that survives contact with 109 power seats. A recommendation you have to walk back costs more than it saved.
Decision
Baseline against you, not the industry
Progress is graded against what this workspace itself did in month one, team by team.
Trade-off No benchmark chart. Buyers ask for one in every first call.
Outcome The number cannot be dismissed as “our company is different”, because it is only ever about their company.
Decision
Publish the gaps
The connector matrix names the vendors that expose nothing, next to the ones that do.
Trade-off We advertise our own blind spots on the marketing site.
Outcome Security review stops being an interrogation. They already read the answer.
Decision
Never read prompts
Metadata only. Where a vendor feed can carry conversation content, it is discarded at ingestion.
Trade-off Content quality scoring is off the table permanently. It is a real feature we cannot build.
Outcome Deployment is a read-only OAuth scope, not a works council negotiation.
User flow / 09
Connect once. Then it keeps answering.
The friction in this category is the first hour. Every step below is designed to remove a reason to stop.
01
Connect accounts
One read-only grant per vendor, from the console you already pay for.
No agent to deploy, so no security review to schedule.
02
Data syncs
Daily pulls. Every vendor's units map into one schema.
Nothing to reconcile by hand, so nothing to get wrong.
03
Review insights
Four figures, each labelled. Then the ranked list underneath.
No configuration first. The baseline is your own history.
04
Take action
Reclaim seats, right-size tiers, route enablement at the team that needs it.
Every item names its seats, so nobody argues about the list.
05
Measure improvement
The same pipeline scores whether it worked, against the same baseline.
The loop closes. That is the part the exports never did.
Outcomes / 10
$4.04k a month back, and one goal that stops missing.
On the modelled workspace, priced line by line. These are expected impacts on a workspace shaped like this one, not results from a deployment.
Spend before
$0.0k
Per month, 287 licensed seats, six tools.
Recovered
$0.00k
0.0% of AI spend. Four line items, zero power seats touched.
Spend after
$0.0k
109 power seats untouched. 74 already on the right tier.
Derived
Recoverable spend
$4.04k a month, 8.4%. Each line traces to named seats. This is the part you can bank.
Modelled
Adoption
Depends on people changing how they work, which is why this page puts no date on it. The product measures whether it happened.
Modelled
Reporting effort
The week of exports becomes a page that is already current. We will not quote an hours figure until a customer gives us one.
Why this instead / 11
The difference is not features. It is whether you can defend the number.
Vendor consoles · a BI dashboard · Depthdata
Measured
| | Vendor consoles | BI dashboard | Depthdata | | -------------------------- | ---------------------------- | ------------------------------ | ------------------------------ | | Tools covered | One each | Whatever you build an ETL for | Eight, one schema | | Definition of “active” | One per vendor, incomparable | Yours, if someone maintains it | One, published | | Baseline | None | Whatever you pick | Your month one, per team | | Spend and value together | No | Only if you model it yourself | Yes, both labelled | | Confidence on each figure | No | No | Measured, derived or modelled | | Estimates ship with a band | No | Rarely | Always. Never a point estimate | | Effort to keep current | A week of exports | An engineer, permanently | A daily read-only sync | | Reads prompts | n/a | n/a | Never. Metadata only |
The row that matters is the fifth. Every product in this table can show you a number. Only one of them tells you how much to trust it.
What this becomes / 12
Where it goes next. Direction, not dates.
No timelines on this page. Coverage depends on what each vendor exposes, and they ship on their schedule rather than ours. The matrix already says so, and a roadmap that pretends otherwise would contradict it.
01
Benchmarking
Your adoption against your own sector, once there are enough workspaces for the comparison to mean anything. Not before.
02
Forecasting
Spend ninety days out, with a band. The same rule as everything else: an estimate arrives labelled, or it does not arrive.
03
Ranked recommendations
The waste ledger already ranks by recoverable dollars. The step after is the action, and the evidence for it, in one place.
04
Agent visibility
Agents are where the spend is heading. Two of the eight tracked vendors offer nothing but a seat list today. When that changes, this follows.
This is direction, not a commitment. Nothing here is sold, dated or promised. If you need one of these to be true before you can buy, say so and we will tell you honestly where it is.
Point it at your own seats
Every figure above
was one workspace. Try yours.Twenty minutes of read-only grants and you will know which of your seats have never been opened, and which of your numbers you are allowed to take to the board.
Get early demo → Read the methodology
✦ FAQ
The questions we get asked first.
✦ CONTACT
Know what your AI is actually worth.
Connect your tools once. Get answers you can take to the board.
READ-ONLY ADMIN APIS · NO AGENTS ON LAPTOPS · PROMPTS NEVER READ
Name\*
Work email\*
Company\*
What are you trying to find out? (optional)
SUBMIT →