Question 1

How is Hobasa different?

Accepted Answer

Five differences in how Hobasa works: it is built for the gaps between systems, reconciling payroll, general ledger and HR data; it keeps one governed Finance + HR model across vendors and subsidiaries; when a rule, law or policy changes it re-evaluates your data and quantifies the exposure; figures come from executed queries and findings are expert-reviewed before release; and it is delivered and configured per client by a named team. Conversational Q&A, dashboards and forecasting are widely available today - Hobasa's difference is doing them on a reconciled, rule-aware, expert-reviewed foundation.

Question 2

How is Hobasa different from BI tools like Tableau, Power BI or Looker?

Accepted Answer

BI tools visualize data you have already modeled. Hobasa connects, reconciles and models Finance + HR data, then monitors and reviews it. Tableau can read your tables; Hobasa tells you where your tables disagree and what it is costing you.

Question 3

We already run Snowflake or Databricks with Genie or Cortex. Why add Hobasa?

Accepted Answer

Keep them - they are the platform. Genie and Cortex answer questions over semantic models your team builds and maintains. Hobasa adds the Finance + HR domain on top: the ontology, cross-system reconciliation, compliance monitoring with quantified consequences, forecasting use cases, and expert review, delivered rather than built by your data team.

Question 4

How is Hobasa different from FP&A tools like Anaplan, Pigment or Adaptive?

Accepted Answer

FP&A tools manage plans. Hobasa monitors actuals across Finance and HR and explains what changed and why. It works alongside planning tools rather than replacing them.

Question 5

How is Hobasa different from people analytics like Visier or Workday?

Accepted Answer

People analytics platforms offer deep HR analytics. Hobasa joins HR data to the general ledger, margin and cash, across mixed vendors - so cross-domain questions get one consistent answer.

Question 6

Why not just use ChatGPT, Claude or Copilot on our exports?

Accepted Answer

They are excellent for one-off analysis, and some now connect to NetSuite, QuickBooks or Gusto. But each session starts over: there is no persistent reconciled model, no monitoring, no compliance monitoring and no review trail. In FinSheet-Bench (2026), an independent academic benchmark, the best model scored 82.4% on financial spreadsheets, and the authors conclude that no standalone model is reliable enough for unsupervised professional finance use. Hobasa uses frontier models inside a governed pipeline where figures come from executed queries.

Question 7

Does Hobasa replace our BI or planning tools?

Accepted Answer

No. Hobasa reads from your systems and works alongside BI and FP&A tools. If your main need is free-form self-serve dashboards or driver-based budgeting, those tools are the right choice.

Question 8

Tableau and ThoughtSpot cost far less per user. Why pay a premium for Hobasa?

Accepted Answer

Because the license prices are not for the same thing. A BI seat pays for the visualization and query layer - the last stages of a much longer journey. Connecting ERP, payroll, HCM and ATS, cleaning and reconciling them against each other, building and maintaining a Finance + HR model, encoding rules and reviewing findings still has to be paid for. In an enterprise that becomes an internal program: data engineers, domain experts, an integrator and months before anyone trusts the numbers. Hobasa includes those stages in the subscription. The right comparison is Hobasa against the program, not against a seat.

Question 9

How do we justify the cost internally?

Accepted Answer

Use four levers, measured on your own data during the first engagement: leakage recovered (unbilled work, duplicate payments), exposure avoided (compliance and rule breaches found early), hours released (analyst time spent reconciling spreadsheets) and decisions made earlier. The first engagement has written acceptance criteria, so value is tested before you expand.

Question 10

What exactly is included, and what costs extra?

Accepted Answer

Included: implementation by a named team, connectors in scope, the configured Hobasa data model, reconciliation, monitoring, alerts, compliance monitoring, forecasting use cases, conversational AI, dashboards and decks, expert review within the agreed allowance, and ongoing support. Quoted separately: additional systems or entities, custom connectors, custom development and extra review capacity. Your order form states the scope.

Question 11

How is Hobasa's output reviewed?

Accepted Answer

There are two kinds of output. Instant answers to questions you ask are computed in seconds from queries on your data, with source records - they are not individually reviewed; you judge them. Reviewed findings that Hobasa surfaces - anomalies, leakage, rule breaches, forecasts - go through expert review before release. For each finding, a Hobasa finance or HR expert checks the evidence behind every figure, the validation results, that totals reconcile to source, that the right rule and threshold were applied, and that the consequence is quantified correctly. Approved findings are released to the right role views and assigned an owner and a KPI to track.

Question 12

How do we know the numbers are right?

Accepted Answer

Four safeguards: figures are computed by queries against reconciled data and carry their source records; the data is validated by multiple layers of checks; the language model is constrained to tool results; and findings are reviewed before release. Quality is measured on your data - reconciliation accuracy, forecast performance against simple baselines, alert false-positive rate - and reported during the engagement. Hobasa does not publish generic accuracy figures.

Question 13

Does the AI hallucinate?

Accepted Answer

Language models can make mistakes, which is why Hobasa does not let the model compute numbers. Figures come from executed queries, the model is instructed to state no figure a tool did not return, forecast answers are checked against published model outputs, and findings are expert-reviewed. Instant answers are not individually reviewed, so treat them as exploration.

Question 14

Is our data safe?

Accepted Answer

Hobasa is ISO 27001 and SOC 2 Type II certified, with read-only access, per-tenant data separation, PII masked before data reaches AI models, SSO with enforced admin MFA, role-based access with audit trails, and customer data never used to train external models. Formal security documentation is available for your review.

Question 15

How fast is it, really? What does "live in 2-4 weeks" mean?

Accepted Answer

For one priority use case, one entity and a small number of systems: the use case runs on connected, reconciled data with signed-off KPI definitions, users can ask questions, monitoring is on, and the first reviewed findings are delivered. The clock starts when read-only access and a named owner are in place. Broader scope is planned separately.

Question 16

What do we need to provide?

Accepted Answer

Read-only system access, a named owner, a short walkthrough of how the business runs, and time from your finance and HR experts to confirm mappings, decide between conflicting definitions and act on findings.

Question 17

Is it customized, or one-size-fits-all?

Accepted Answer

Configured per client: your KPI formulas and terminology, systems and subsidiaries, rules and thresholds, industry pack and role views. Customization lives in configuration, not in a fork, so every client receives every product improvement.

Question 18

What is available today, and what is roadmap?

Accepted Answer

Available today: connectors, reconciliation, validation, the knowledge graph, conversational AI, monitoring, compliance monitoring, seven forecasting use cases, expert review, dashboards and decks. Roadmap: scenario bands, stage-specialized small language models, time-series foundation models, named expert sign-off on each model release, and integration with ticketing tools.

Frequently asked questions.

How and why Hobasa is different

Value and price

Trust and review

Delivery

Have a question we did not answer?