What a data specialist told us firms miss before AI

AI is only as good as the data underneath it. A firm's data has to be organised, complete and secure before any AI goes on top, because AI works quickly and confidently from whatever gaps are there. Louisa Stewart of Blue Ninja is the person who goes in and finds those gaps, and this autumn we had two long conversations with her.

Blue Ninja helps small and growing businesses with their data, automation and information workflows. Louisa has worked with data across many sectors, and much of her work happens inside a client's own systems, checking what is really there against what people believe is there.

Firms are skipping the data to get to the AI

Louisa sees her clients rush past the data. "They kind of skip over the details to get into the AI," she said, "and it's like, but you haven't got all of the process sorted in order for the AI to sit on top of it." A fortnight later she was hearing it everywhere: "I keep hearing more and more people say, 'We're losing sight of important data strategy and quality,' in order to try to speed up the AI side of it."

We work in the other order. Before any solution is planned, we check the layers it will sit on: the business and each department, the tech stack and whether its software can connect to AI, the data, and the people. For the data, that means its quality, its completeness, who owns it and how secure it is. Where a layer isn't ready, fixing it goes first on the roadmap.

A firm often holds less of its data than it thinks

Checking the data means going to the source. Louisa doesn't rely only on system data exports, because "the system's going to show you what it wants to show you." At one conference, the organisers believed they controlled their attendee data, because the company collecting it sent over a spreadsheet. Louisa compared the form with the spreadsheet and told them, "you're missing three different fields here."

A firm is in the same position with every system that holds its client records. AI can only use data the firm can reach, and the firm can only change tools if its data comes with it, so we want its data and knowledge held where it controls them. Then it isn't tied to one AI vendor: if a provider raises its prices or changes its terms, a different model can work from the same material.

Gaps in old records cost time on every job

Missing fields also gather inside a firm's own records over time. Someone once asked Louisa why they had no city for most of their contacts when they were "collecting the cities". The field had only been added to the form a week before, and "around 60,000 contacts" had been added to the system without it. As she put it, "at what point does the data then become validated?"

In a firm, gaps like that cost money before any AI is involved. Every job has a time budget, and the fee is set from it. When a client's records are incomplete, someone goes back for what's missing, the job runs over, and the extra time is written off, so the firm bills less of the time it recorded and its recovery rate falls. A solution reading a firm's client history runs into every gap in it, so we check the data across the whole period a solution will read from. Fixing those gaps goes on the roadmap with its cost, next to what the solution returns. Once the jobs it touches come in inside their budget, less is written off, and the recovery rate, fee income and gross profit rise together, with the same team taking on more work without adding headcount.

To know how much of your team's week goes on repeat work and how often jobs run over budget, and what to do about it in the next 30 days, take the quiz below. It takes a few minutes.

Complete data needs an owner in every department

Keeping the data complete is a job for the whole firm. "Often with the departments, they only know their own stuff," Louisa said, and "it also comes down to the teams taking responsibility for putting data in and making sure that they're collecting the right data. And the data's complete when they do it." The same goes for where data ends up: "if someone puts a confidential financial report into their own AI, there's no control over that."

So every solution we implement has a named owner who answers for what it produces and for the data under it staying complete, and who checks its output before it reaches a client, so the standard of work holds or rises. Where it runs depends on the data it touches: the AI provider's own cloud when nothing sensitive is involved, the firm's own server where the provider never trains on what it sends, or a machine in the office. The firm's AI policy covers it, including any AI the team already uses on personal accounts. As the models change, the owner keeps checking the solution and tweaks it before its quality slips without anyone noticing.

What firm leaders should do now

Louisa was honest that data is a hard sell. "Clients will always say that they don't have time for data," she said, and in her experience they "only see costs in data." That changes once the data work belongs to a solution the firm wants, with its cost set against what the solution returns. Taken together, her points set out the order a firm needs to follow, whoever does the work.

Start with the firm's own work, broken down into the individual tasks and processes that create value for its clients: who does each one, how long it takes, what data it uses and who signs it off.

Before any solution, check the layers those tasks sit on: the business and each department, the tech stack, the quality, completeness and security of the data, compliance and governance, and the people who will work alongside the AI. Where the data isn't organised or complete, fixing it comes first.

Then set out a roadmap of solutions, each fully costed with the hours it frees up, what it costs to implement, its first-year return and what it does to the recovery rate on the jobs it touches, with the data work each one needs costed alongside it and, where possible, done in the same phase, so the firm sees results while its data improves.

Implement the solutions one by one, in the roadmap's order. Each runs where its data allows, has its risks designed out, is covered by the AI policy and has a named owner who checks what it produces before it reaches a client. Once they are live, keep managing them as the models change, with the firm owning its data and knowledge and renting the AI.

Done in that order, the firm gets more capacity from the same team, hits its chargeable time targets and lifts its recovery rate and fee income, in a way that is compliant and keeps client data secure.

If you want us to do it for you, book a free call and we'll find you at least $50,000 a year in new business that AI can unlock. It's compliant, data-secure and has governance built in.

Know exactly where your team's hours go, and what to fix first

Our quiz asks how much of your team's week goes on repeat work, how often jobs run over budget, how your team uses AI, where client data goes and how you would know whether AI paid off. Your report shows where AI would free up your team's time first, the capacity it could give your firm in hours, and a 30-day plan for what to do.

We'll find you $50,000 a year in new business AI can unlock in 72 hours, or you don't pay.