Financial Model Review
WealthAi Financial Model Review
Advisor Briefing for the WealthAi Team
Executive Summary
A rebuilt, driver-based financial model is now ready to sit alongside the March 2026 plan. It starts from first principles (sales funnel, cohort conversion, tiered pricing, cost base, expansion curves) and produces a full set of statements. Where it aligns with the existing model, that is validation. Where it diverges, it is usually because the rebuilt model adds mechanics the spreadsheet could not capture: expansion, churn, scenarios, and a working cash flow.
Headline: with a GBP 3M seed closing June 2026, the business has a tight pre-seed runway of 8.5 months from April, drops to 6.2 months at the seed close, then resets to 16 months runway post-close. Year-end 2027 is the second tight point at 6.8 months runway before 2028 profitability lifts the business clear. Net revenue retention is 125% in 2027 and 108% in 2028 on a cohort basis.
About This Model
The Excel workbook is now formula-native. Every number is calculated live in Excel from the Assumptions tab, with a parallel Python implementation powering the web dashboard. The two are cross-validated to zero percent diff across 24 check points. The Assumptions tab is the single source for every driver: change a value there and the whole workbook recalculates.
Engine (Python)
- All drivers, formulas and assumptions live in code
- Versioned in git, every change is diffable
- Covered by unit tests on drivers and engine
- Three scenarios (Conservative, Base, Aggressive) from the same source of truth
- To change an assumption, edit the driver inputs and re-run
Excel (Output)
- Read-only presentation of engine results
- 13 sheets: Summary, Intro, Key Drivers, Funnel, Revenue, Costs, Hiring, P&L, Balance Sheet, Cash Flow, Unit Economics, Valuation, Reference
- Not interactive: changing a cell does not recalculate
- Regenerated in seconds from the engine when assumptions change
This is more rigorous than a pure spreadsheet: the logic is testable, versioned, and repeatable, and there is no risk of a broken formula silently changing an answer. If investors want to stress-test an assumption, we change one input in the driver config and regenerate the entire workbook.
What This Model Adds
The existing March 2026 plan is a client-by-client revenue schedule. Every named client is listed with a tier, contract value and start month, and the monthly MRR builds from that list. It is detailed and grounded in real pipeline data. The rebuilt model keeps that pipeline story intact and adds the mechanics investors will ask about.
Existing Model (March 2026 plan)
- Individual named clients on a monthly schedule
- Full blended TCV from day one for every client
- No churn modelled
- No funnel or conversion mechanics
- No unit economics (CAC, LTV, NRR)
- Formula errors on the P&L and cash sheets
- Single scenario
- No balance sheet or statement of cash flows
Rebuilt Model (v3)
- Driver-based: funnel → conversion → clients → revenue
- Land-and-expand modelled for every tier (85-100%)
- 10% annual churn modelled
- 8 driver categories on one Key Drivers sheet
- Full unit economics: CAC, LTV, NRR, Rule of 40
- Working P&L, Balance Sheet and Cash Flow
- Three scenarios: Conservative, Base, Aggressive
- March 2026 plan preserved on a Reference sheet for side-by-side comparison
How the Numbers Compare
| Metric | March 2026 Plan | Rebuilt Model | Delta |
|---|---|---|---|
| Clients (year-end) | |||
| Dec 2026 | 12 | 11 | -8% |
| Dec 2027 | 44 | 50 | +14% |
| Dec 2028 | 96 | 116 | +21% |
| ARR (year-end) | |||
| Dec 2026 | £1,036,900 | £731,945 | -29% |
| Dec 2027 | £7,212,770 | £5,747,540 | -20% |
| Dec 2028 | £13,871,740 | £15,475,087 | +12% |
| Expenses (annual) | |||
| 2026 | £1,978,510 | £1,870,498 | -5% |
| 2027 | £3,705,922 | £3,844,697 | +4% |
| 2028 | £4,820,497 | £4,911,837 | +2% |
The shape of the story is consistent across both models: a slow 2026 build, a steep 2027 ramp, and a profitable 2028. The differences are explainable and, in most cases, make the rebuilt model more defensible to a diligent investor.
The rebuilt model shows more clients at lower initial ACVs in 2026 and 2027 because large clients are assumed to land at a fraction of their full tier TCV and expand upward over the first 6 to 18 months (see next section). By 2028 the cohorts have matured, expansion has compounded, and total ARR pulls ahead of the existing plan. Expenses track within 5% across every year, which is strong validation of the cost base.
Key Findings
1. The land-and-expand story is now visible in the numbers
This is the single biggest change from the March 2026 plan and it deserves to be the headline. The rebuilt model lands every client at a fraction of their tier TCV and ramps them to full contract over the first phase of the relationship:
- Boutique lands at 85% of TCV, expands to 100% by month 6
- Small lands at 80%, expands to 100% by month 9
- Medium lands at 70%, expands to 100% by month 12
- Large lands at 60%, expands to 100% by month 15
- Very Large lands at 55%, expands to 100% by month 18
This matches what is actually happening in the pipeline (Plurimi landing at GBP 35K on a GBP 196K tier, Mediobanca landing at GBP 50K on a GBP 495K tier) and produces net revenue retention of 125% in 2027 and 108% in 2028. That is the compounding platform thesis made numerical: land with compliance, expand into research and advisory.
This is the investor story
NRR above 120% is what separates a compounding SaaS platform from a collection of point solutions. Every board slide, investor conversation and KPI dashboard should lead with expansion revenue alongside new logos. The cohort programme produces the proof points. If expansion runs faster than modelled, the upside is significant. If it stalls, the Conservative scenario shows the downside.
2. Cash runway is comfortable with a GBP 3M seed
This was tight in earlier versions of the model. With the raise sized at GBP 3M and closing in June 2026, cash stays positive across the entire plan and never drops below roughly GBP 1.0M. The low point is mid-2027, after which revenue begins to fund the business.
| Milestone | Cash Balance | Comment |
|---|---|---|
| Apr 2026 (model start) | £479K | Opening balance |
| Jun 2026 (pre-close trough) | £230K | Final pre-seed month |
| Jul 2026 (seed close) | ~£3.2M | +£3M seed |
| Dec 2026 | £1,384K | Cash positive |
| Aug 2027 (low point) | ~£1,021K | Low point of the plan |
| Dec 2027 | £2,054K | Ramp has taken over |
| Dec 2028 | £10,304K | Strong position for Series A |
The GBP 3M raise does the job without forcing an emergency bridge in 2027. There is still no fat in the plan, but the model shows the business reaching EBITDA profitability during 2028 from this single round.
3. Unit economics are healthy across all three years
| Metric | 2026 | 2027 | 2028 | Benchmark |
|---|---|---|---|---|
| CAC | £56,503 | £57,908 | £48,166 | £40-80K vertical SaaS norm |
| Entry ACV (Year 1) | £62,737 | £106,172 | £108,416 | |
| Mature blended ACV | £64,102 | £114,261 | £132,941 | |
| LTV:CAC | 8.3x | 13.8x | 16.9x | 3-5x good, >5x excellent |
| CAC Payback | 14.4 months | 8.7 months | 7.1 months | 12-18 months good |
| Gross Logo Retention | 90% | 90% | 90% | 92-95% target |
| NRR (cohort) | n/a | 125% | 108% | 110-120% good |
| Revenue / Employee | £10K | £116K | £329K | £150-300K good |
Capital-efficient, defensible unit economics
CAC includes Commercial and Marketing payroll fully loaded with 30% employer cost loading, 10% sales commission on new ARR, and a 25% sales loading uplift covering ramp inefficiency, accelerators and enablement tools. LTV is calculated on entry-discounted ACV (not mature blended), so expansion is not double-counted with NRR. Year 1 CAC payback at 14.4 months sits inside the 12-18 month industry target. LTV:CAC of 8.3x in year one is in the excellent band, easing into 13-17x as the platform compounds. Gross logo retention is reported separately so each metric is investor-defensible.
Modelled, not yet observed. The metrics above are derived from the financial model V2 driver engine. The first cohort is in flight. Observed CAC and ACV will be measurable after Q3 2026 close. NRR will be measurable after 12 months of usage at the anchor client (Q4 2026). Until then, treat these numbers as forward-looking targets that we are committed to validating.
4. Valuation benchmarks have been reset to defensible comps
| Multiple | Dec 2026 | Dec 2027 | Dec 2028 |
|---|---|---|---|
| 10x ARR (Conservative) | £7.3M | £57.5M | £154.8M |
| 15x ARR (Base) | £11.0M | £86.2M | £232.1M |
| 20x ARR (Aggressive) | £14.6M | £115.0M | £309.5M |
The existing plan uses 24x ARR across the board. That sits above the comparable transaction range (Unique 15-20x, Nevis ~15x, Avantos 12-15x at Series A) and is difficult to defend in diligence. The rebuilt model uses 10x/15x/20x as the three scenario multiples, which is the standard range for vertical AI SaaS at seed.
At 15x on 2027 ARR, the business is worth approximately GBP 86M pre-Series A. That is a strong outcome for a seed round at this stage and leaves clear upside for the next round.
5. Things that still need confirmation from the team
Named pipeline assumption
The 2027 ARR ramp relies on specific named clients converting at specific times (Mediobanca, Plurimi, Rathbones, JM Finn among others). The rebuilt model uses a funnel-led approach that assumes this pipeline holds. If any of these deals slip out of 2027, the profile shifts closer to the Conservative scenario. Please confirm the named pipeline is still live.
Hiring plan and actual payroll
The rebuilt model has 28 roles across Management, Technology, Product, Commercial, Marketing and Legal. The executive team has more roles in flight than are currently modelled. Sharing the actual monthly payroll and planned hire dates will let us tighten the cost base. Expenses already reconcile to within 5% of the existing plan, so this is about precision rather than structural change.
International ramp timing
The model currently assumes Italy and Switzerland open in Q4 2026, with each geography taking roughly 6 months to reach steady funnel throughput. Either the model timing moves to 2027, or the public positioning explicitly carves out the model assumption as more aggressive than the conservative external narrative.
Expansion velocity
The NRR story depends on clients actually moving from compliance to research and advisory. The Launch Group cohort is the evidence base. If expansion happens faster than the curves above, every scenario gets stronger. If it stalls, NRR drops toward 90% and the Conservative case becomes the realistic one. This should be the KPI the team tracks hardest from Q3 2026.
Three Scenarios
| Metric (Dec 2028) | Conservative | Base | Aggressive |
|---|---|---|---|
| ARR | £5.0M | £15.5M | £27.7M |
| Active Clients | 68 | 116 | 213 |
| EBITDA | -£2.7M | +£4.0M | +£12.3M |
| Implied EV (scenario multiple) | £50M (10x) | £232M (15x) | £554M (20x) |
The Conservative case is the stress test: expansion stalls, large clients never graduate, churn eats into the base. Even here the business reaches GBP 5M ARR with a manageable burn. The Aggressive case is what happens if the cohort model scales faster than expected and the international ramp lands on time. The Base case sits on the current product, team, and pipeline and is the number to plan against.
What Needs to Happen Next
Before investor meetings
- Confirm the named 2027 pipeline. Mediobanca, Plurimi, Rathbones, JM Finn and the rest. Which are still in, which are at risk, what is the honest probability on each. The ramp depends on it.
- Share the real payroll numbers. The cost base is close, but tightening it with actual salaries and start dates takes one more objection off the table.
- Reconcile the international timing. The model still has Italy and Switzerland opening Q4 2026. Canonical positioning is 2027 for both. Pick one and propagate, or explicitly carve out the model assumption as more aggressive than the public narrative.
- Agree the expansion narrative. NRR of 125% in 2027 is the story. The team needs a consistent pitch on what drives it (cohort graduates, new modules, geographic replication) and what the early proof points are.
After seed close
- Track actual vs model monthly. The Summary sheet has a month-by-month view. Compare actual MRR, client count and burn from month one.
- NRR is the north star. Expansion revenue from the first cohort graduates is the single metric that proves the platform thesis. Build the KPI dashboard around it.
- International readiness gate. Italy and Switzerland hires only fire if the UK playbook is documented and repeatable. Whatever launch date is chosen (model says Q4 2026, canonical says 2027), the gate is the same: the UK playbook has to work first.
Deliverables
Driver engine (Python)
Source of truth for every number in this briefing. Lives at model/ with unit tests in test_drivers.py and test_engine.py. Three scenarios defined in config.py.
Excel workbook (generated output)
13-sheet presentation at model/output/WealthAi_Financial_Model.xlsx. Sheets: Summary, Intro, Key Drivers, Sales Funnel, Revenue, Costs, Hiring Plan, P&L, Balance Sheet, Cash Flow, Unit Economics, Valuation, Reference. Regenerated from the engine whenever assumptions change.
Reference sheet
The March 2026 plan data is preserved inside the workbook for side-by-side comparison. Nothing has been lost, and every difference between the two models is traceable.