AI Data Analyst At Your Sevice
The AI Analysis module reads your cost-KPI tracking data and writes a brief executive summary in seconds, surfacing the key concerns, opportunities, and anomalies worth your attention. From now on, you always know how your money is spent, and what it produces.
The AI model never performs any of the underlying calculations — it only narrates results that have already been computed by the analysis engine.
The AI Analysis module is included for every client, at no additional cost.
Instant Executive Brief About Cost-KPI Performance
Select any date range, and the AI Analysis module reads the underlying cost-KPI results for that period and writes a plain-language summary: which cost categories show the strongest connection to your results, where the data flags a concern, and where there’s an opportunity worth acting on. It reads like a briefing a senior analyst would prepare before a board meeting, ready in seconds instead of hours.
You get the same depth of interpretation every time, regardless of how many cost categories you’re tracking or how complex the underlying data is. Instead of working through charts and tables yourself, the module surfaces what actually needs your attention first, so nothing important gets buried under the numbers that don’t.
FAQs
Does the AI Analysis module run automatically?
No. You request an analysis manually from a dedicated tab, selecting the date range you want to examine. You can run separate analyses for costs, KPIs, or both combined. This keeps the module on demand rather than generating unsolicited output every time you add data.
What data does the AI model actually see?
It reads your full dataset for the selected period — cost entries, KPI figures, and the statistical results already computed by the analysis engine. Cost entries are structured as a category label, an amount, and a reference code. No company name, supplier name, or contextual information beyond that is included. The model sees numbers and references, not your business identity.
Which AI provider processes the analysis?
You can connect your own API key from any supported provider — OpenAI, Anthropic, or others. If you do not add a key, the platform defaults to Anthropic. Your data goes to whichever provider you have configured, and is subject to that provider’s standard API data handling terms.
Can the AI find things the charts wouldn't show me?
Yes, and this is one of the more practical uses of the module. Because it reads the raw input data alongside the computed results, it can flag things like duplicate invoice references, irregular payment frequency, cost categories that appear inconsistently across periods, and anomalies that would be easy to overlook when reviewing charts manually.
Does the AI perform any of the statistical calculations?
No. All calculations — the regression analysis, Pearson correlations, Spend Efficiency Score, and Cost Flow results — are computed by the platform’s own analysis engine using standard statistical methods. The AI model reads those results and writes a plain-language interpretation. It narrates; it does not calculate.
How much data do I need before the analysis is useful?
The module will run analysis on whatever data you have entered, but results become meaningfully more reliable as your dataset grows. With fewer data points, patterns are harder to distinguish from noise. Ten weeks or more of consistent weekly entries is where the analysis typically starts producing conclusions you can act on with confidence. During onboarding you will have the option to backfill historical data, which is worth doing if you have past records available.
Can I trust an AI to interpret financial data accurately?
The module is not making financial judgements — it is describing statistical outputs that have already been computed. It tells you what the numbers show, not what you should do about them. The underlying calculations are deterministic and auditable. The AI’s role is to surface what matters and write it in plain language, so you spend less time reading tables and more time making decisions.
Is my data used to train AI models?
API usage through major providers such as Anthropic and OpenAI does not use submitted data for model training under their standard commercial API terms. If you use your own API key, the data handling is governed by your agreement with that provider. We recommend reviewing your chosen provider’s API data policy if this is a compliance consideration for your organisation.
What does an analysis output actually look like?
It reads as a short executive brief — the kind a senior analyst might prepare before a board meeting. It identifies which cost categories show the strongest connection to your results, where the data flags a concern, where an opportunity appears, and whether anything in the raw entries warrants a closer look. It is written in plain language, not statistical terminology, and is typically a few paragraphs.
What if I only started using the platform recently?
You are not limited to the data entered since signup. The platform allows you to backfill historical cost and KPI data, and this is actively encouraged during the onboarding process. The more historical context the analysis has, the more reliable the patterns it identifies. Starting with six to twelve months of past data, where available, will produce a substantially more useful first analysis than starting from a clean slate.