Roiva Docs

Answered from these docs only, by a model that cannot see your account. Check the pages it cites.

Product Pricing Sign in Start free

Building an AI ROI tracker yourself

Starting free, when vendor review blocks that, what a build takes, from cost pipelines to signed attribution rules and time studies, and where it breaks.

View as Markdown

Answered from these docs only, by a model that cannot see your account. Check the pages it cites.

Checked against the product on October 12, 2026

You can build this yourself. The method Roiva uses isn't proprietary. It's published in these docs and on the methodology page, and an analyst can turn it into tables, formulas and an approval flow.

This page covers where to start, what a build and its upkeep take, and where home-built trackers tend to fail. It also says which of those failures Roiva handles and which it leaves with you.

Start in the free account

If the plan is a lean tracker now and a build-or-buy decision later, the lean tracker can be Roiva's free account. It's permanent, needs no card, and holds:

  • Your whole portfolio, with 3 initiatives measured at a time, and 10 people. Ideas you haven't started don't count, so the full list fits. The people cover preparers, a backup, the approvers who sign figures off, and a reviewer from finance or risk.
  • Figures you type in or import from a CSV: costs, metric readings and baselines, assumptions and value entries. A spreadsheet of initiatives imports in one step.
  • The controls a spreadsheet can only ask for. Value counts only once a person approves it, and by default nobody can approve an entry they recorded, last edited or entered a figure it was calculated from. Each calculated figure keeps the formula version and inputs it was made with. Closing a month freezes its costs and approved value until someone reopens it with a reason, and every change is in the audit log.
  • Every template, with its formulas, metric definitions and assumption defaults, so you don't design them from scratch.
  • Measurement charters and time studies. Each initiative's attribution rule, written down and signed by someone other than its author, and the before-figures timed with the people who do the work. Neither is held back for a paid plan.
  • Up to 3 connections, enough for one initiative's value source, its vendor bill and its AI provider usage to fill themselves.

What it doesn't do is measure more at once or connect every system: a fourth measured initiative, a fourth connection, the reporting tables for Power BI and your warehouse, and the read API. That's what a paid plan adds. Whatever you put in the free account stays when you add any of it, so the build-or-buy decision becomes a decision about the plumbing, made on a record you already have.

To get in, enter your work email at roiva.ai/sign_up. Roiva emails you a link that works for 24 hours, and following it lets you choose a password and name your account. Or choose Continue with Google or Continue with Microsoft on the sign-in page with a work account. A personal address, such as Gmail, gets in by taking the assessment on the same page. If your organization's Roiva account has verified your email's domain, a confirmed work email or a Google or Microsoft sign-up takes you there instead, to sign in or ask its admins for an invitation. If you've been invited to an account, the email links to the invitation. You can also request an invite, and we set the account up within one business day. See Plans and billing for the limits.

If vendor review blocks even a free account

A regulated team often can't put company data into a new vendor until its risk process has reviewed that vendor, free or not. For that, an Owner can turn on Aggregate figures only. The account then holds only what your people type in, import, or confirm from a Roiva suggestion. Roiva refuses every connection, keeps no files, and sends nothing to an AI model or to product analytics. An Owner can also allow cost connectors, which bring in only your cloud and AI-provider bills and usage totals. Aggregate figures only: a note for your risk team says exactly what such an account stores and who else sees it, and is written for you to hand to that team.

When a spreadsheet is enough

If your team can't use a vendor at all yet, or your portfolio is small, a spreadsheet or a few tables in a warehouse you already run is a reasonable answer. Roughly, that means up to five initiatives and two or three spend sources, plus someone other than each initiative's owner who will sign off its figures every month. At that size you can check for double-counted cost by eye, a formula change is a note beside the cell, and one person can hold the whole thing in their head.

The method matters more than the tool. Whatever you build, keep these rules:

  • Only approved value counts, and every cost counts whether anyone approved it or not.
  • Report value minus cost, never value on its own.
  • Keep cash apart from freed capacity. Hours saved at a labor rate aren't money the business stopped spending, so report them beside cash, take ROI over cash alone, and move capacity into cash only when a document shows it became cash.
  • Store each figure with the version of the formula that produced it and the inputs it used, so it can be re-run later and come out the same.
  • Record how each baseline was measured, and keep an estimate labeled as an estimate. A before-figure measured after go-live, on work done without the AI, is a reconstruction, and should be counted apart from one measured before it.
  • Where no system records how long the work took, time it with the people who do it. If the study times the work both with the AI and without it, draw each person's days without it at random rather than letting anyone choose, work out the interval by person rather than by timing, and set a date to time it again.
  • Write down what the AI gets credit for, and what it doesn't, and have someone other than its author sign it before the results are in. A change to that rule after results needs a reason, recorded where the next report shows it.
  • Whoever prepared a figure doesn't approve it.
  • A reported month doesn't change unless someone reopens it and records why.

Roiva's tracker template is a workbook laid out on these rules. Its Summary sheet adds up approved cash value against every cost, with freed capacity beside it and never added, and its Checks sheet counts the rows that break a rule, since Excel can show a broken rule but can't refuse one. Kept as it's laid out, the workbook imports into a Roiva account later in one step. Initiatives, costs, baselines, readings, assumption history, value entries with their inputs, who approved each one and when, closed months, and time studies with their timings all come in. Your Excel formulas don't: each figure stays the figure the workbook had. See Importing a tracker workbook.

The breakpoint is less about the number of initiatives than about sources and people. Once the same vendor shows up in more than one system, or initiatives change owners, the spreadsheet stops being a record and becomes something one person maintains.

What a build takes

A tracker that a finance team will accept, for a portfolio of about a dozen initiatives, needs these pieces:

  • Cost pipelines from each provider's billing API or export, your ERP's vendor bills and your card program, with rules for when the same spend arrives from two of them.
  • Metric pulls from the systems that hold the before and after figures: the support desk, the CRM, source control, and your own operational data.
  • Versioned formulas with input snapshots. Each monthly figure stores which version of the arithmetic produced it and a copy of every input it used.
  • An insert-only audit record. A status change is a new row, and no role can update or delete one.
  • An approval flow that records who approved each figure and refuses an approval by the figure's preparer.
  • Period locks, so a reported month can't change without a recorded reopening.
  • Reporting that reads only approved value, with cash and freed capacity as separate figures, and split by whether each figure rests on a measured, reconstructed or estimated before-figure.
  • A signed attribution rule for each initiative, frozen when it's signed, with a check that flags a figure calculated some other way and a record of any revision made after results.
  • Time-study tooling, wherever no system holds the before-figure: how many timings and people each study needs, a random draw of who works without the AI on which days, a way for people outside the finance team to record minutes, an interval worked out by person, and a date each result is due again.

A rough estimate for that portfolio, with six or seven spend sources at a company that already runs a data warehouse and a BI tool, is 12 to 16 person-weeks to build, spread over two to three months. Most of that goes into the cost pipelines and their de-duplication rules. It doesn't include the charter check or the time-study tooling: a signed document and a shared sheet can stand in for them, without the draw, the interval or the check. Running it then takes about 0.3 to 0.5 of a full-time person. That covers gathering and approving figures each month across owners and finance, keeping the pipelines working, and repeating time studies once or twice a year.

Treat those numbers as estimates, not measurements. Yours depend mostly on how consistently your cloud spend is tagged and how clean your ERP's vendor data already is.

Where home-built trackers break

None of these is certain. They're the failures that tend to appear once a tracker is a year old and the person who built it has moved on.

  • Approver independence erodes. An owner approves their own figure "just this once" because the approver is on leave, or approval becomes a monthly batch nobody reads. Unless the tool refuses it, the rule only exists in a policy document.
  • Baselines get overwritten in place. Someone corrects a baseline cell, and every figure calculated from it changes, including last quarter's. A reported number that moves without a record is what makes a CFO stop trusting the rest.
  • Cost gets counted twice, or three times. A model provider's API bill, the same vendor's invoice in your ERP and a card charge for the same subscription can be one dollar recorded three ways. Cloud invoices are harder still: an AWS or Microsoft invoice mixes AI and non-AI services, so it can't be assigned to an initiative whole. De-duplication rules are the fiddliest part of a build, and a new vendor or card program can break them without raising an error.
  • Tags aren't retroactive. A cost-allocation tag applies to usage from the moment it was put on the resource. Spend from before then carries no tag, and it can only be split by estimate.
  • Key-person risk. A home-built tracker usually lives in one analyst's head. When that person leaves, the next one inherits pipelines, formulas and exceptions they didn't write.
  • Vendor APIs change. GitHub shut down its legacy Copilot Metrics API on April 2, 2026 (GitHub's notice), and the endpoints that replaced it report different fields. A pipeline built on the old API either failed, or, if it swallowed the error, went on reporting success while returning nothing. Every provider you pull from makes changes like this on its own schedule.
  • The rules move after the results. A disappointing quarter is followed by a higher credit share or a corrected baseline, and nothing records that the rules changed after the numbers were seen. A CFO who finds one such change discounts the rest.
  • Time studies that don't survive a question. The person timing chooses which tasks go in the before arm, or forty timings from two people are reported as a tight average. The first draws the comparison, the second overstates how sure it is: two people's timings say how fast those two work.
  • "Append-only" becomes a claim you have to prove. Insert-only tables and tight database grants make a sound record. An auditor will still ask you to show that no role can change a row, and that nobody used the one that can. That becomes a controls document you write and keep current.

What Roiva takes off your plate

  • Cost from the platforms it connects to, with double-counting handled. Where a vendor's name belongs to that vendor alone, its own report is counted, and the copy on your books is left out in any month the report is booked, except the seats and other products the report doesn't carry, which still count from your books. Card charges aren't counted a second time from the card's account in your ledger. For AWS, Azure and Google Cloud, both copies are booked and the overlap is reported rather than guessed at. Anything Roiva can only suspect is listed, not dropped. Finance → AI Spend says how much may be counted twice, and each vendor's page shows the copies. See How costs reach an initiative and What each platform books.
  • Metrics that a connected platform reports sync on their own, and baselines are calculated from the months before an initiative started where those months are available. The metrics reference lists which metrics a connection can feed.
  • Metrics you already model in a warehouse. A claims cycle time or a quote count that lives in Snowflake, Databricks, BigQuery or SQL Server is read by a metric query you write once. Roiva versions the query, keeps the rows each version returned, and runs it every night. See Metrics from your own data warehouse.
  • Getting the figures back into your warehouse and BI tool. On a paid plan, once an Owner or Admin turns them on, Roiva builds its figures every night as flat, typed tables: approved value with the formula version and inputs behind each figure, costs, approvals, budgets against spend and decisions. It checks them against its own reports to the cent before anyone can read them. It can write them into an S3 bucket you own, for Snowflake, Databricks or Fabric to load, and Power BI can read them directly. Your warehouse stays the one place your reporting reads from. See Roiva's figures in Power BI and your data warehouse.
  • Measurement charters. Each initiative's charter holds its attribution rule in your words, with the formulas, shares and baselines it runs, frozen when it's submitted and signed by someone other than its drafter. A figure calculated outside it is marked Outside its charter and never approved automatically, and a revision after value was approved needs a reason and is printed on the Board Brief. See Measurement charters.
  • Time studies. Roiva plans how many timings each study needs, draws each person's days without the AI, and gives anyone outside the account a link of their own with a timer. It imports a CSV of timings exported from Zendesk, Freshdesk, Harvest, Toggl or Jira, works out the interval by person, and publishes what was timed without the AI as the baseline and what was timed with it as a reading. It also reminds the study's owner when the result is due again. A study of agent minutes per ticket also prices the tickets your AI agent deflects. See Running a time study.
  • Baselines for an initiative already live. The months before go-live, a sample timed without the AI, or a declared estimate, each recorded by how it was measured and reported apart. See Baselines for an initiative that's already live.
  • Formula versions and input snapshots. Every calculated figure keeps the version and inputs it was made with, and a later calculation leaves an approved month alone. Correcting a baseline next quarter doesn't rewrite a figure someone already approved. See What makes a figure auditable.
  • Approver independence, enforced by the app. By default nobody can approve a value entry they recorded, last edited or entered a figure it was calculated from. The rule steps aside only when nobody else could approve, and only the account Owner can turn it off under Organization Settings → Approvals, which leaves a record in the audit log. See Approving value and initiatives.
  • Period locks, on any plan and with or without a connection. Locking a month freezes every cost entry dated in it, however it was recorded, and every approved value entry, until someone reopens it with a reason. Once a QuickBooks, Xero or NetSuite connection books bills through an allocation rule, locking can also post journals to your books. See Allocation rules and accounting periods.
  • Upkeep of the connections. When a provider changes its API, updating the integration is Roiva's work rather than yours, done once for every account instead of once per company.

What Roiva leaves with you

  • The timing itself. Roiva plans a time study, draws the days and works out the result, but your people time the work. Whether the tasks they timed stand for the rest, and whether they worked at their usual pace, is a judgment no tool makes.
  • Approvals. A person still reads each figure and decides whether to sign it off. Auto-approval exists and is off by default, and even when it's on it holds back any figure that rests on a default assumption, is above your approval threshold, or comes from a formula whose figures no person has approved before.
  • Metrics nothing holds yet. A figure that is in no connected system and no data warehouse is still recorded by hand or imported from a CSV. That is the same work a home-built tracker needs. Writing the metric query for a figure your warehouse does hold is yours too, though it's one SELECT per metric rather than a pipeline.
  • The attribution rule. How much of a change the AI gets credit for is your judgment. Roiva records the rule in a signed charter and holds every figure to it. It doesn't test whether the change would have happened anyway, and no before-and-after comparison can.
  • Tagging at the source. Roiva can split cloud spend by tag, workspace or project, but only if those tags, workspaces and projects exist. Spend from before you set them up is as hard to split in Roiva as anywhere else.
  • Deciding suspected duplicates. When a typed cost and a synced bill might be the same money, Roiva lists the pair and a person decides which copy counts.
  • Your own vendor review. Roiva doesn't have a SOC 2 report yet. Its SOC 2 Type 2 program with Oneleet started on September 25, 2026. The examination covers the Security criteria and includes a three-month observation period, and the same engagement includes a penetration test of the application every year. Until there's a report, the security page says what you can ask for, including Oneleet's engagement letter, and Roiva's policies and reports are requested through the Roiva Trust Center.

Choosing

  • If you can use a free SaaS account, start in Roiva's free account rather than a spreadsheet. You get the controls without building them, and the record is already where the plumbing would plug in if you later want it.
  • If vendor review is the blocker, start in the free account with Aggregate figures only on, and turn it off once the review clears.
  • If neither is possible yet, keep the record in the tracker template, or a spreadsheet of your own built on the rules above. Revisit the decision when a fourth spend source appears or the person who built it moves on. A record kept in the template imports into Roiva when you're ready.

If the pipelines, the de-duplication and the audit trail are the parts you'd rather not own, that's the part Roiva does. The judgment, the approvals and the timing stay yours either way.