Helping a deep-tech CFO plan systems before the scramble

Client: Anonymised UK deep-tech company (early stage, raising and selling)
Their side: CFO
What we did: Discovery call, then a written systems roadmap (discussion draft — not a priced project)
Our side: SH Applications — Steven Hallissey FCA

No names. No product detail you could Google. This is about how we work with a finance leader in a serious tech business — not a write-up of their confidential pack.

The situation

They build hard technology, not a simple app. Selling is technical. Deals often start as trials or pilots. The team is still small. Money is planned carefully around a fundraise path.

A new CFO had kept finance deliberately simple: online accounts with an outside bookkeeper, forecasts in the CFO’s head and spreadsheets, and not much shared system yet for “who’s buying, what are we delivering, what did it cost?”

That is fine before real customers pay. It gets painful the moment paid work starts stacking up — different customers, different projects, more people asking “can we afford this?”

The real question was not “which AI tool should we buy?” It was:

What do we need in place so we can sell, deliver, invoice and forecast without chaos — and without ripping out the accounts system mid-fundraise?

What we did not do

• Open with a product list or a day-rate quote
• Tell them to replace their accounts package
• Promise a fixed six-month project before the picture was clear
• Make up numbers we had not seen

Same rule as an AI Diagnostic: understand the business first, then decide what is worth building.

One structured conversation

We spent about an hour on a plain discovery call: what they do, how they sell, what they already use for money and ops, and what “good enough” looks like before the next big investor conversation.

We agreed a few simple rules:

1. Keep the current accounts system for now — no big finance system swap while raising money
2. One place for customers and deals — sales and delivery detail lives there; clean numbers feed the accounts
3. Small team, smart automation — less re-typing and chasing; people spend time on judgment
4. Numbers you can act on — not a report you have to untangle first
5. Investor-ready means real control — pipeline, delivery cost, cash and runway you can explain calmly
6. Stay flexible on how they charge — pricing models will evolve; the systems should not need a rebuild every time

What they got: a roadmap, not a quote

A few days later we sent a discussion draft — a clear written plan plus a short slide pack for leadership. In plain terms it covered:

• Where they are — stage, pressures, what already works
• Gaps and risks — especially as paying customers arrive
• Target picture — how sales, delivery, accounts and forecasting should fit together
• First three months — foundations: cash clarity, customer/deal tracking, clean invoicing, a forecast the CFO trusts each week
• Next three months — deeper project costing, more automation, a steadier board/investor rhythm — with checkpoints, not a locked plan
• Money view — short-term cash, full-year forecast, deal economics, simple scenarios
• Whole business, not only finance — sales, delivery, people, risk and IT as well
• How we’d work together — meeting rhythm, open actions, risks

No price in the pack. Cost depends on how much is hands-on advice versus things we can build once and reuse. That talk only starts if they say they want to go ahead — then NDA and a proper engagement, not a surprise invoice on page one.

How this links to putting a £ on opportunities

A systems roadmap for a CFO is close cousin to an AI Diagnostic. Both force the same habits:

• Name the friction (copy-paste, thin project tracking, forecast cut off from real work)
• Put a practical value on fixing it — see How we put a £ on AI opportunities
• Rank do now / do next / do later so the first months are foundations, not a wish list

Here, “do now” was not a fancy dashboard. It was the basics that make first invoices, project margin and cash runway visible without heroics.

What “good” meant for this stage

Success at this point is clarity, not a finished build:

• The CFO has a written map to test with the CEO and board — not a vague “we should get a CRM”
• The rules are clear (keep accounts; track customers properly; automate the boring bits)
• The first three months have a sharp goal: sell, deliver, bill and forecast without chaos
• The next phase has gates — we only go further if the foundations work
• The commercial next step is simple: do you want to do this? → yes → NDA + engagement

Whether they hired us to build it is between us and them. The public lesson is the method.

Why other founders and CFOs should care

If you are early-stage and heading toward a bigger raise, the usual failure is not “we didn’t have a chatbot.” It is:

• First paying work lands in a mess of spreadsheets
• The forecast lives in one person’s file while the team works somewhere else
• Investor questions expose process theatre instead of real control

The answer is not a longer software shopping list. It is a ranked plan, written in plain English, with a finance brain and enough tech sense to know what to automate first.

In one line

Listen like a CFO. Map like an operator. Rank like an investor. Build only what the ranking justifies.

That is the same idea as our AI Diagnostic — used here for systems and forecasting in a deep-tech business.

Client identity and materials are confidential. This article is anonymised and describes approach only.