Every consultancy owner has had the month where a client cancels, a big invoice slips 45 days, and payroll is due Friday — and the only real question is whether there's enough cash sitting somewhere to not panic about it.
Most operators can tell you their monthly recurring revenue to the dollar. Far fewer can tell you how many weeks they could survive with zero new revenue starting today. That number — not MRR, not close rate, not lead volume — is the one that determines whether a bad quarter is an inconvenience or an existential threat.
This isn't about seasonal cash-flow smoothing (covered separately — that's about timing revenue troughs you can predict). This is about the reserve that sits underneath all of that: the buffer you build once and don't touch unless things actually go wrong.
Why Consultancy Owners Underfund Their Reserve
Service businesses — including AI consultancies built on retainer or licensing revenue — have a specific blind spot: revenue looks recurring, so it feels safer than it is. But "recurring" revenue that depends on 8-12 client relationships is nothing like the recurring revenue of a SaaS company with thousands of accounts. Lose two clients in one month and a consultancy can drop 25-40% of revenue overnight. A diversified subscription business barely notices losing two customers.
That mismatch — feeling like recurring, behaving like concentrated — is why so many operators run lean on cash. They see stable-looking numbers and reinvest instead of reserving. Then a single bad month becomes a scramble: a missed vendor payment, a delayed hire, or worse, a rushed decision to discount a renewal just to keep cash moving.
The Three-Number Framework
Instead of a vague "save more" goal, three concrete numbers make reserve planning actionable.
1. Fixed monthly burn. Payroll, software/vendor stack, rent or coworking, insurance, loan payments, your own minimum draw. Not revenue-dependent spend — the number that doesn't change if every client vanished tomorrow.
2. Runway target in months. How many months of fixed burn should sit in reserve before you touch anything beyond it. This scales with client concentration, not company age.
3. Trigger threshold. The reserve balance that, if crossed, forces a defined response (pause hiring, pause non-essential tool spend, revisit pricing) rather than a panicked one-off decision made under stress.
Runway Target by Client Concentration
| Client Concentration | Recommended Runway | Why | |---|---|---| | 1-3 clients (any one client = 30%+ of revenue) | 6 months | Losing one client is a revenue cliff, not a dip | | 4-8 clients | 4 months | Meaningful but survivable single-client loss | | 9-15 clients | 3 months | Loss of any one client is a manageable percentage hit | | 16+ clients, no client above 15% of revenue | 2-3 months | True diversification reduces the tail risk reserve has to cover |
An operator with three clients licensing ConsultancyOS-style infrastructure and one of them worth 40% of revenue is carrying more concentration risk than an operator with twelve smaller accounts — even if their total revenue is identical. The reserve target should reflect that, not a generic "3 months of expenses" rule of thumb pulled from personal finance content.
Where the Reserve Actually Lives
A reserve that isn't separated from the operating account isn't a reserve — it's just a bigger number that's easy to spend down without noticing. Three practical structures work:
- Separate business savings account, no debit card attached, manual transfer required to move funds out. Friction is the point.
- Two-account split: operating account funds day-to-day spend, reserve account only receives a scheduled monthly transfer and is never the default account for outgoing payments.
- Short-term treasury or money market account for reserves beyond 3 months of burn, so the excess isn't sitting idle at 0.01% while still being accessible within a day or two if genuinely needed.
The mistake to avoid: treating a line of credit as a reserve. Credit is fine as a secondary buffer, but it isn't a reserve — it's debt you'll owe with interest exactly when revenue is already down, which compounds the problem instead of absorbing it.
Building the Reserve Without Starving Growth
Most operators don't lack the will to reserve cash — they lack a rule that survives a good month. The fix is a fixed percentage taken off the top before anything else gets allocated, not a "whatever's left over" approach.
| Approach | How It Works | Failure Mode | |---|---|---| | "Save what's left" | Reserve gets whatever remains after all spending | Almost always ends at $0 — spending expands to meet available cash | | Fixed % off gross revenue (5-10%) | Reserve transfer happens automatically before other allocations | Works consistently; requires discipline to not "borrow" from it | | Milestone-triggered (e.g., save 100% of next new client's first month) | One-time boosts tied to specific wins | Good supplement, unreliable as the sole method — depends on new sales happening | | Windfall-only (bonus months, big one-off projects) | Reserve grows only from unplanned upside | Reserve growth becomes unpredictable and often never happens |
The fixed-percentage approach is the one that actually gets reserves built, because it removes the monthly decision. Automating the transfer the same day revenue lands — before it's "available" to spend — does more than any spreadsheet discipline.
When the Reserve Should Actually Get Used
A reserve that's too precious to touch is functionally useless. The trigger threshold from the framework above exists so "should we dip into it" isn't a debate every time cash gets tight. Decide in advance what qualifies:
- Yes, use it: a client cancels with no notice and payroll is due before a replacement pipeline closes; a major vendor or licensing payment is due before an expected receivable lands; a one-time compliance or legal cost (an E&O claim excess, for example) needs to be covered immediately.
- No, don't use it: funding a new hire speculatively; covering a founder's discretionary spend increase; absorbing a client's slow payment as a permanent pattern instead of fixing the collections process that let it happen.
The second list matters as much as the first. Reserves get quietly drained by "temporary" uses that become permanent, and by the time an operator notices, the buffer that was supposed to cover a real emergency is gone for reasons that had nothing to do with an emergency.
What This Doesn't Fix
A cash reserve buys time. It does not fix a business model that structurally can't sustain itself — if margins are too thin to ever generate a surplus, no amount of discipline turns that around; the fix is pricing or cost structure, not a bigger buffer. It doesn't replace an E&O or business insurance policy for liability exposure (a separate line item, covered here). It doesn't substitute for actually managing client concentration risk — a healthy reserve makes losing one client survivable, but it doesn't make losing three in the same quarter painless if the underlying client base was always too concentrated. And it isn't a reason to skip contract terms that protect cash flow in the first place, like notice periods and payment terms — a strong contract prevents the cash crunch a reserve is designed to survive.
A Simple Starting Sequence
For an operator with no reserve today, the order that works in practice:
- Calculate fixed monthly burn (be honest — include your own minimum draw, not zero).
- Set the runway target using the concentration table above.
- Open a separate account, no card attached.
- Set an automatic transfer of 5-10% of gross revenue on the day it lands.
- Write down the trigger threshold and the "yes/no" use list before an emergency happens, not during one.
- Revisit the target every time client count or concentration changes materially — a reserve sized for three clients is undersized once you're at ten, even though the dollar comfort might feel similar.
None of this requires guessing about vendor and tool-stack costs or waiting for perfect visibility into client lifetime value — the fixed-burn number is knowable today, and the runway math follows directly from it.
FAQ
How much cash reserve does an AI consultancy actually need? Use fixed monthly burn × the runway target from the concentration table, not a flat industry number — a 3-client shop and a 15-client shop have very different risk profiles even at identical revenue.
Should the reserve include personal savings or only business funds? Keep them separate and size each independently. Business reserve covers business burn; personal reserve covers personal expenses. Mixing them hides how exposed either side actually is.
Is a business line of credit a substitute for a cash reserve? No. A line of credit is a secondary buffer, not a reserve — it adds interest cost exactly when revenue is already down, which is the opposite of what a reserve is for.
How does client concentration affect the runway number? Directly. The fewer clients you have, and the larger any single client's share of revenue, the more months of runway are needed to survive losing one — see the concentration table above.
Where This Fits Into Building a Durable Operation
Cash reserve planning is one piece of running a consultancy that can absorb a bad month without an existential scramble — alongside structured client reviews that catch churn risk early and a realistic view of what building vs. licensing AI infrastructure actually costs to run. Operators licensing ScaleLogix AI's ConsultancyOS get the infrastructure and lead pipeline built for them, but the financial discipline around reserves, concentration risk, and runway is still the operator's own responsibility — no licensing program replaces having your own buffer in place before you need it.
If you're evaluating whether a licensed AI consulting model or a fully self-built stack fits your risk tolerance better, see if you qualify for a conversation about what the numbers actually look like for your situation — reserve planning included.