Letting Small Business Owners Ask Their Forecast What Happens Next
40% to 50% shorter budget cycles,
the 2026 benchmark where AI-assisted planning replaces rebuilding the model by hand
60% to 70% faster forecast updates,
with the same class of tooling, which is the loop this agent shortens to a single sentence
Real numbers, never guesswork,
Webpuppies connected the assistant to the platform's own forecasting engine, so every figure comes from the owner's actual books rather than from the AI's general knowledge
About the Project
Services
Project Details
The client is an Australian platform that helps small business owners understand their money. Tradespeople, service businesses, cafes, shops. It plugs into the owner’s accounting software, turns their trading history into a forecast, and tracks how the year is actually going against it, month by month.
It starts from an unusual question. Rather than asking what revenue the business should chase, it asks how much money the owner needs the business to make to pay for the life they want, then works backwards to how they might get there. Owners can try out an idea and watch what it does to profit and cash before committing to anything.
The Challenge
The platform already did the hard part well. The forecasts were accurate, the connection to the accounting software was clean, and the monthly tracking worked. What kept getting in the way was the gap between having good numbers and being able to ask them anything.
Financial software is very good at showing what has happened. Asking it a question is a different skill. An owner who wants to know what happens if they hire someone in March, put prices up, or take on a bigger job first has to turn that into the shape the software expects. Most people running a small business do not think in that shape, and they should not have to.
What the Next Level Required
Questions asked the way people speak
Answers from their books, not the internet
Told what it means, not just what it is
Something their accountant can check
The platform needed an assistant built into the product and wired to its own forecasting engine, not a chatbot bolted onto the side of it. That meant building on top of the numbers the platform already trusted.
Our Solution
An Assistant Built Into the Forecast, Not Bolted Onto the Side of It
Building AI Into a Financial Planning Product
Webpuppies built an assistant that reads the owner’s forecast, their accounts and how the year has actually gone, then answers questions from all three. It works inside the platform’s own sums, so a question about hiring is answered against that owner’s real cash position rather than a general rule about staff costs.
Ask It the Way You Would Ask a Person
Owners ask about profit, cash, margins or a particular month in plain words, and the assistant works out which parts of the forecast the question touches. The step that used to stop people before they started is simply gone.
Try an Idea in a Sentence
Testing something now takes one line. The assistant applies the change, shows what happens to profit and cash, and lets the owner tweak the idea and ask again. Weighing up three options no longer means building three forecasts.
It Tells You Why, Not Only What
The assistant explains why a number moved and what that means for the months ahead, in the language a business owner uses. That is the difference between a figure on a screen and a decision someone is willing to make.
Your Accountant Can Check the Working
Every answer points back to the forecast lines and the months behind it. The owner can see how it was worked out, and an accountant looking at the same account sees exactly the same basis.
Ongoing Partnership
An assistant like this gets better as real questions pile up. Webpuppies keeps working with the platform on what owners actually ask and where the answers need to be sharper.
More Kinds of Question
Widening what the assistant handles confidently as patterns in real use emerge.
Bigger What-Ifs
Handling several changes at once, the way real decisions actually arrive.
Sensible Running Costs
Keeping the cost per question proportionate as more owners come on board.
A View for the Accountant
Building the advisor side, so owner and accountant discuss the same reasoning.
Results & Impact
Forecasting that answers back
The platform went from showing numbers to talking about them, and owners who would never open a scenario builder will happily ask a question. The figures above are published 2026 industry benchmarks for this kind of tool rather than measured results from this engagement, and they show what changes when trying out an idea stops meaning building a forecast from scratch.
Related Projects
MyRepublic
Increasing Conversions with a Region-Specific, User-Friendly Website
See how we helped MyRepublic create an intuitive, multi-regional website that boosted engagement and conversions.
Ron Kaufman
Merging Two Websites into One Seamless Platform
Find out how we helped Ron Kaufman, one of the world’s leading customer service keynote speakers, integrate his two websites into one seamless platform for a better user experience.
