FAQ
Everything you'd want to know before letting AI anywhere near your finance function — data, security, sign-off, pricing, and rollout.
What Becide is, and what it isn't.
All of the above, deliberately combined. Becide is an AI services company for finance — Business Avengers' team delivers the service (data cleaning, configuration, ongoing support), running on a platform (the Becide data lake and AI layer), which shows up to you as the tools you use day to day (the Management Report, AI Assistant, Voice Agents). You're not left to configure a generic tool yourself — you're engaging an AI services partner.
An AI services engagement for your finance function, built on your cleaned data: the Management Report, the AI Assistant, and the AI Voice Agent for AR follow-ups, all live together. Broader capabilities (consolidation, predictive analytics, CRM/procurement connections) are roadmap additions to the same services offering, not separate products.
Broader AI for finance — that's the whole point. The Management Report, the AI Assistant, and Voice Agents are three equal, live capabilities today. Management reporting tends to be the easiest thing to demo concretely, but it isn't the identity of the company.
No. Becide provides analysis and reporting to support decision-making; outputs should be reviewed by the client's team and a qualified finance professional for anything with legal or significant financial effect.
Business Avengers Intelligence Labs, founded by three co-founders: Dilip Kherajani (CPA, 20+ years, including years as a practising CFO), Rohit Maloo (Chartered Accountant, MBA Finance), and Aditya Shah.
The honest comparison against Claude, ChatGPT, and friends.
Uploading a spreadsheet into Claude gives you a one-off answer on whatever's in it — errors included, no memory between sessions, no audit trail, and no guarantee your data stays confined to that conversation. Becide is a standing platform: your data is cleaned once by people who do this professionally, structured into a governed data lake that persists and improves over time, and every answer is logged and cited.
With a general AI tool, your data is processed under that vendor's own infrastructure and terms, which may not guarantee Australian data residency. With Becide, everything — your data, the data lake, and the AI processing itself — runs inside AWS's Sydney region under a dedicated encryption key unique to your organisation, and it's never used to train any model.
The step every other AI vendor skips.
That's expected, not a blocker — it's exactly what Business Avengers' onboarding exists to fix. The team reviews your existing systems and data directly, correcting errors, gaps, and misclassifications by hand before Becide is deployed on top of it.
Reconciling inconsistent source systems is part of what Business Avengers does during onboarding, before Becide is deployed — the AI itself isn't asked to paper over unreconciled data; the data is fixed first so the AI has something reliable to work from.
How Becide gets to your numbers.
Live data connections are available for well-known, major accounting platforms as standard. For anything else, Business Avengers builds a custom connection based on what the data integration assessment finds for your specific systems. CSV/Excel upload is also always available regardless of connector status.
Major, well-known platforms are supported via existing connectors; less common or specialised systems are handled as custom integrations, scoped during the data integration assessment.
Yes — either through custom data extraction built during onboarding, or via scheduled CSV/Excel exports from the source system. The AI itself always runs on the cleaned data in your Sydney data lake, regardless of where the source lives.
CRM and procurement connections are on the roadmap. Today, non-financial data can be brought in via custom connectors or file uploads if it's needed to enrich your management report.
What the first three months look like.
Business Avengers does, working with you during onboarding — KPIs are identified per client based on industry and what your board actually tracks, not left for you to configure yourself in a generic dashboard builder.
No, deliberately. Becide is built on a data lake that Business Avengers cleans and structures for your business first; it's not a tool you sign up for and self-configure on day one.
That's the point of the model — most of the implementation work (data assessment, cleaning, structuring, configuration) is done by Business Avengers' team, not yours. Your team's involvement is mainly providing access and context up front, and reviewing the configured output before go-live.
Not as a strict requirement — the data cleaning and structuring work is done by Business Avengers. Having someone who can grant system access and answer questions about how your data is organised speeds things up, but a dedicated IT resource isn't a prerequisite.
Yes — the onboarding period is essentially a parallel-build phase by design. You keep running your current process while Becide is being configured, and switch over when you're confident in the output.
Configured per client, not a template.
No — the underlying report structure is consistent, but the specific KPIs, and any agents built, are configured per client.
Yes — repeat-customer rate matters more for a subscription or travel business, cost-of-goods-sold detail matters more for a manufacturer.
Yes — structured around your segments, KPIs, and existing reporting format rather than a generic template.
The core report structure (P&L, variance, cash flow, forecast) applies across industries; the specific KPIs layered on top are tailored per client. Manufacturing, professional services, e-commerce, and subscription businesses are patterns already reflected in current positioning.
KPI customisation is designed to support this — subscription businesses are explicitly a case where KPIs like repeat-customer rate matter more. Confirm the exact metrics in scope during your data integration assessment.
Australian-first architecture. Full technical detail in section 12 below.
Entirely within AWS's Sydney, Australia region — nothing is processed offshore.
AES-256 encryption at rest with a dedicated key per client, TLS in transit, strict role-based access, and a read-only AI that can never modify or delete your data.
A tenant-isolated data lake with raw, cleaned, and reporting/curated layers, all within AWS Sydney; the AI reads only the cleaned and reporting layers, never your original raw files.
What Becide can and can't do today.
Yes — this is the core, shipping capability. Board-ready pack generated directly from your cleaned Data Vault.
Not yet — flagged as a high-value roadmap item, not a current feature. See the multi-entity section below.
Yes, via the AI Assistant, with citations back to the exact source figure behind every answer.
Yes, via the AI Voice Agent — upload a receivables list and script, the agent calls, confirms payment dates, logs outcomes, and always discloses it's AI.
Yes — reports are downloadable after generation. PowerPoint export is on the roadmap.
Yes — role-based access within your organisation, plus Business Avengers' admin access for delivery and support. Fine-grained permissions can be scoped during onboarding.
Yes — the AI Assistant is designed for plain-language follow-up questions in the same thread, with every response citing the underlying figures.
No dedicated mobile app today — Becide is a web platform that works on any modern browser.
The Alerts feature surfaces anomalies and action items automatically within the platform. External notification channels (email/Slack) can be scoped during onboarding.
How Becide sits alongside the systems you already have.
Because an ERP manages transactions; it doesn't automatically produce a board-ready management pack or answer ad-hoc questions with citations. Even ERP vendors bundling their own AI dashboards face the same underlying issue — a dashboard alone doesn't fix data entered incorrectly at the source.
Sitting on top of it — Becide doesn't replace your ERP or general ledger; it's the reporting and insight layer built on your cleaned data, alongside your existing systems.
An in-house analyst still has to manually pull and reconcile data from your systems before they can analyse anything — much of their time goes into assembly, not judgment. Becide automates the assembly step so a human spends their time on judgment instead.
Becide isn't a substitute for CFO-level judgment and isn't positioned as one — it's what a CFO (yours, or Business Avengers') uses to spend less time producing reports and more time on strategy.
Becide is currently in pilot. Here's the honest state of it.
Becide is currently in pilot. You pay for the data cleaning and data-lake build service; the Becide platform itself is free for the first three months while the product is refined based on real use. Ongoing pricing after the pilot period hasn't been finalised — get in touch for what's currently on offer.
No — nothing is metered per query, report, or token. The current model is a fee for the data-cleaning/build service, with the platform itself free during the three-month pilot.
Not billed separately — during the pilot, platform use (including report generation) is included at no extra charge.
Not currently — no metered cap, prepaid credit, or auto-top-up model is in place during the pilot.
Not yet decided — pricing beyond the current pilot structure (data-cleaning fee plus three months free platform access) isn't finalised.
Yes — this is the current model. Every onboarding client gets the Becide platform free for the first three months, paying only for the data cleaning/data-lake build. This isn't a limited-feature trial; it's the real product, priced as pilot access.
The upfront fee covers the data cleaning and data-lake build, which unlocks free platform access (Management Report, AI Assistant, Voice Agents) for three months. Ongoing fee structure after that is being finalised now.
Not yet decided — pricing beyond the current pilot structure is still being finalised.
Governed by the Refund Policy — see that page for the confirmed terms.
Against dashboards, BI tools, and category-standard FP&A software.
Tools like Flowcast and Trintech add analytics and visualisation on top of whatever data is connected — if the underlying data is inconsistent, the dashboard displays that inconsistency faster. Becide's data has already been cleaned and structured by Business Avengers before a report is ever generated.
A dashboard is table stakes at this point. The actual bottleneck for most finance teams isn't visualisation, it's the days it takes to assemble an accurate pack in the first place. Becide's differentiator is producing that pack in minutes, from data verified clean by Business Avengers.
These tools generally connect to accounting platforms and visualise the data as-is; the same "we clean your data first" distinction applies. Becide sits above the connector-plus-dashboard pattern by putting people, not just software, on the cleaning step.
The direct trust questions a skeptical CFO has about AI touching their numbers.
By construction: the AI only ever answers from your own cleaned, structured data lake (never general internet knowledge), every card and figure carries a clickable information icon showing exactly where the number came from, and the AI's access is strictly read-only so it can't quietly alter the underlying numbers. Users can also flag and correct any response, feeding the correction back into the agent's memory.
Every card and figure on the platform carries an information icon a user can click to see exactly where the data came from — the underlying source and the path it took through the pipeline — so accuracy is checkable per figure, not just asserted.
Yes. Every AI response has a feedback mechanism — you can flag a response and provide a correction, and that feedback is stored in the agent's memory so future responses reflect it.
"Large language models" — the specific underlying model is an implementation detail likely to change over time. What's fixed is the architecture around the model: read-only access, per-client isolation, citations on every output, and the human review step before any number is treated as final.
The infosec questionnaire, answered.
Becide is built on AWS, hosted in the Sydney, Australia region — a cloud platform widely used by Australian government agencies and major financial institutions. All client data is stored and processed entirely within Australia.
Yes, always. Every component — file storage, database, data processing, and the AI itself — runs entirely within AWS's Sydney region. Client data never leaves Australian jurisdiction at any point.
No. Becide's entire infrastructure, including the AI system itself, runs inside AWS in the Sydney region.
Access is strictly role-based. Each client's data is encrypted with its own dedicated key, so it's technically impossible for one client's data to be read under another client's access. Multi-factor authentication is enforced for all administrative accounts.
All stored data is encrypted at rest using AES-256. Each client has a dedicated encryption key held inside AWS's key management service (KMS) — the key never leaves AWS and is only used at the instant data is processed. Data in transit is fully encrypted. Login passwords are one-way hashed; database connection passwords are AES-256 encrypted, never logged, and unlocked only for the instant a connection opens.
No — not even internal platform administrators. Only the system itself can unlock a key at the moment of processing.
Three layers, all within AWS Sydney: original files exactly as uploaded (encrypted with the client's own key, untouched, accessible only to the client); a cleaned and structured layer used for analytics and AI queries; and a reporting/command-centre layer. The AI only ever operates on the second and third layers — never the original raw files.
PII is treated as sensitive throughout. It's encrypted at rest with the client's dedicated key, access is role-restricted, and it's never sent to third-party AI training pipelines. Where possible, PII is masked or minimised before being included in AI-facing layers.
Every AI run — every report generated, every question answered, every administrative action — is logged onshore with what was checked, what tools were used, and what the outcome was. Logs are retained per your agreement and available for review.
You may export your data within the period stated in your agreement. After export or expiry of the restoration window, we delete data from all layers and schedule destruction of your encryption key — ciphertext becomes unreadable, including in backups.
No. Client data is never used to train third-party AI models. Period.
Australian obligations, honestly stated.
Yes — Becide is operated in accordance with the Privacy Act 1988 (Cth) and the Australian Privacy Principles. Our Privacy Policy sets out how personal information is collected, used, and disclosed.
Not currently held as third-party certifications. Our controls (AES-256 encryption, per-client keys, tenant isolation, read-only AI, audit logging, onshore data residency) are documented and available for enterprise security review as part of onboarding.
Yes — a countersigned DPA is available on request. See the DPA page for the standing template.
We will notify affected controllers without undue delay after becoming aware of a personal information breach affecting their data, providing information reasonably available to support their assessment and notification obligations under the Privacy Act.
Where multi-entity support stands today.
Not as a fully automated consolidation engine yet — this is a high-value roadmap item. Today, multiple entities can be onboarded and reported on separately, with cross-entity summaries handled during the data structuring step.
Yes — separate entities can each be modelled in the Data Vault, and Becide can generate a management report per entity. Automated group-level consolidation is a roadmap item.
What happens after you say "yes, let's talk."
There's no self-serve signup — deliberately. Onboarding starts with a real conversation about your data, followed by a data integration assessment. Use the "Talk to us" form to start.
A short intro conversation about your finance function, then access to (or exports from) your accounting/ERP system for the data integration assessment. Nothing more upfront.
The assessment conversation can usually happen within a week of first contact. From kickoff, a live configured Becide instance is typically three months out.
How to verify what we're claiming.
Becide is in pilot, so a public case-study library isn't published yet. Enterprise prospects can request references on a case-by-case basis during the sales conversation.
Business Avengers Intelligence Labs — the labs arm of Business Avengers, an Australian outsourced accounting, bookkeeping, and CFO advisory firm with 12+ years of experience serving Australian businesses.
Who you're contracting with.
Becide is provided by Business Avengers Intelligence Labs, part of the Business Avengers group. The specific contracting entity and ABN are set out in your order form or engagement letter.
The laws of New South Wales, Australia, unless otherwise agreed in your order form. See the Terms page for full detail.
The odds and ends that don't fit above.
Not currently — Becide is offered as a direct engagement between Business Avengers and the end client. Partner/reseller models can be discussed on a case-by-case basis.
Yes — onboarding includes hands-on walkthroughs for your team on the Management Report, AI Assistant, and any Voice Agents built for you. Additional training sessions can be scoped.
Your data remains yours. Contractual arrangements include export rights and data-portability commitments — see the DPA and Terms pages for detail.
The bigger picture behind the product.
Because most "AI for finance" tools were built somewhere else, on infrastructure that routes data through jurisdictions Australian CFOs and boards can't easily audit. When your data crosses a border, so does your ability to enforce Australian privacy law on it. Keeping everything in Sydney removes that entire class of question.
General AI became useful for finance work about two years ago. The gap that opened up isn't "can AI answer a finance question" — it clearly can. The gap is trust: can you put it in front of a board without a person to verify every number? Becide is built for the current window, where the AI is capable enough and finance teams are ready to use it, but only if the trust architecture around it is real.
Still have a question?
Email us or talk to our team. Real answers, no sales script.