Systems integration and the finance judgment behind it — your ERP, Salesforce and the connections between them, the reporting layer, and the three-statement model the numbers feed. When the numbers are late it is rarely a people problem; the machinery was never built for where you are now. We rebuild it, and we own the outcome.
Every one of these came out of a real engagement. Click any that land — each one opens to show what is actually causing it, and the panel underneath reads the pattern across everything you pick.
Most CEOs select between five and nine of these. The interesting part isn’t how many — it’s which cluster they fall into, because that tells you whether you have a data problem, a cash problem, a systems problem, or a decision-making problem wearing three different costumes.
This is not a menu you assemble from. It’s a single engagement that owns the outcome — which in practice means most of the work is technical, because that’s where the problem lives. The finance judgment decides what to build. The engineering makes it exist.
Sales lives in Salesforce. The books live in the ERP. In most growing companies the space between them is a person with a spreadsheet — which is where bookings and billings stop agreeing, customers get created twice, and reps call finance to ask whether an invoice was paid.
We design the hand-offs once: which system owns which record, what flows in which direction, and who gets alerted when something fails. Then we build it and keep it running.
We map where every number actually originates, decide which system is authoritative for each, and build the connections so they stop disagreeing. Revenue means one thing. A customer is one record. The number in the board deck is the number in the ledger, because it came from there.
Payments, billing, e-commerce, 3PL, payroll, spend — every one of them either reconciles to the ledger or generates manual work forever. Pre-built connector where one genuinely fits, custom API where it doesn’t, and the part most integrations skip: monitoring, alerting and an error queue with a named owner, so a failure surfaces in minutes instead of at month-end.
Years of custom scripts, workflows, fields and reports nobody can explain — or an implementation that went live and never quite worked. We inventory what’s there, retire what’s dead, fix what’s broken, document what stays, and put testing around releases so the next upgrade doesn’t break month-end.
An integrated P&L, balance sheet and cash flow that actually tie — so a hiring plan, a price change, an acquisition or a covenant test gets run as a scenario instead of argued from instinct. Cut three ways: defensible line by line for the bank, driver-level for the board, and operable weekly by the managers who move the numbers.
The dashboard that’s been “almost ready” for two quarters is usually not a BI problem — it’s that the data underneath was never made to agree. We fix that first, then build the warehouse and reporting layer on top of numbers that reconcile. Then it just refreshes.
Approval routing, invoice capture, reconciliation matching, close checklists — automation that removes a human step without removing the control.
Then the part almost nobody has done yet. A company brain: how your business actually works, written into structured, version-controlled files a model can read — policies, close procedures, pricing logic, contract terms, and the reasoning behind decisions you already made. Retrieval runs against that instead of against a folder of PDFs, so “why do we do it this way” survives the person who knew.
Open-weight models on your own hardware or inside your own cloud tenancy where the data cannot leave. Agent harnesses wired into the ERP and the CRM under real permissions — processes that read and write, not a chat window bolted onto the side.
AWS, GCP and Azure architecture and migration. Single sign-on and access control an auditor will accept. Backup and recovery that has actually been tested. Most of what a lender’s IT questionnaire or a SOC 2 asks for is architecture, decided long before anyone sends the questionnaire.
A revolver or an ABL to fund working capital. Venture debt to buy quarters without dilution. Acquisition financing. A priced round. Or funding the next phase from your own cash generation. The deliverable is a forecast that can be defended line by line and shared with your board and your bank.
We’ll run the month-end close, manage the accounting and tax vendors, keep the reconciliations clean and handle the audit. It’s table stakes and we don’t lead with it — but it does get done properly, and you won’t have to think about it again.
Most firms in this space are either one photograph and one CV, or an IT shop that stops at the network. Both top out the moment the work crosses between systems and finance — which it always does.
ERP, Salesforce, the data platform, and the connections between them. Engineers who have run this in production, not advisors who write a recommendation and leave.
The model, the forecast, the capital strategy, the board and bank material. The person who owns the number is the person who built it.
Custom internal tools, agent harnesses, migrations, the thing no vendor sells because it is specific to how you operate.
Releases, user access, enhancements and the error queue — handled by people who already know your environment. Not a helpdesk, and never a ticket that sits for a week.
We assemble against the problem rather than staffing a pyramid — and we are deliberately selective, because this only works on engagements where the ambition is worth the horsepower.
These are the moments that force the issue. In every one of them the finance question and the systems question turn out to be the same question — which is why they tend to get handed to two different vendors and then fall down the gap between them.
Salesforce runs the pipeline, the ERP runs the books, and a person with a spreadsheet runs the space between them. Every quarter-end, bookings and billings disagree and someone spends a week finding out why — usually a deal that changed after it closed, or a customer that exists twice.
An investor, a lender, or a first auditor. The findings are always the same shape: control gaps that are really system gaps, deferred revenue maintained by hand, no audit trail on journal entries. Getting ready means changing the systems, not writing a memo about them.
Two entities, two charts of accounts, two ERPs, two identity providers, and a board expecting consolidated numbers inside thirty days. The synergy number in the memo depends entirely on integration work that nobody has scoped, costed, or assigned to anyone.
Close takes eleven days. Finance spends the month reconciling rather than analyzing. Every real question becomes an export and a pivot table. It stopped fitting two or three years back, and nobody has had a spare quarter since to do anything about it.
AWS or GCP spend grew faster than revenue and nobody can attribute it to a product, a customer or a team. Engineering says it’s necessary, finance says it’s a line item, and both are arguing without data. It’s a gross margin question that can only be answered with tagging, architecture and a cost model working together.
Institutional knowledge walked out of the building. The model is a black box with someone else’s assumptions in it, half the close is undocumented, and nobody remembers why the integration was built the way it was. You need coverage now, not after a four-month search.
A subsidiary, a first foreign entity, or an e-commerce or marketplace channel. A modern ERP handles all three — multi-entity, multi-currency, commerce connectors — but only if the chart of accounts, item master and integrations were designed for it. Retrofitting after launch is the expensive way to find out they weren’t.
Not a hard question either — payback period by channel, or margin on the largest account. You said you’d follow up. Two weeks later three people had built three versions and you presented the average. That is a data architecture problem being experienced as a credibility problem.
Money spent, go-live slipped twice, the partner stopped answering thirty days after cutover, and now nobody internally wants to raise it. The second attempt is harder politically and easier technically, because at least this time everyone knows what the requirements actually were.
Every deal that crosses from Salesforce into the ERP by hand costs minutes, and a share of them go out wrong and cost an hour to unwind. None of it shows up as a line item — it shows up as a person who is always busy. Push the inputs and watch for the month the manual process needs its own full-time hire.
The first month is deliberately front-loaded. If we can’t tell you something useful and slightly uncomfortable about your own business inside three weeks, the engagement was a bad idea and you should know that early.
Read-only access to the ledger, the CRM, the billing system, the bank and the cloud accounts. We reconcile what they each claim and find out where they disagree.
A map of every system and every hand-off between them, a health check on the ERP, and a written read on what is actually wrong — ranked by what it costs you rather than by how hard it is to fix.
The integrations, the close, the reporting layer, the ERP decision — whatever the findings said mattered most. Sequenced so each piece delivers on its own rather than in one big reveal.
Integration monitoring, ERP administration and releases, reporting — and finance work as it comes up. Every artifact documented and yours, so the eventual full-time hire inherits something operable.
There is a version of this work that keeps the books tidy, answers questions when they’re asked, and produces a pack at month-end that nobody opens. It’s honest work and there are good people who do it well. It isn’t what this is.
We take engagements where finance is supposed to make the company better — where something has to be decided, rebuilt, defended or funded. The measure we hold ourselves to is whether the business is harder to knock over, and the team sharper, than when we arrived.
Serious engagements, serious results. And still simpler — and cheaper — than hiring the wrong person full time.
An engagement like this is right for a specific window, and that window closes. Most firms won’t tell you where the edge is, because the edge is where their revenue stops. Here’s ours — if two or more are true, you should be hiring, and we’ll help you scope the role and sit in on the interviews.
The work needs a senior hand more days than not — the constraint is availability, not capability.
Finance has grown past three people and needs a manager who is present daily.
The board wants a named executive with equity alignment in the seat.
You’re heading into a sale process where continuity of a full-time officer matters to buyers.
The systems are stable, the close is boring, and the interesting problems are all commercial now.
Not a discovery call to qualify you — a straight conversation about what isn’t working and whether we’re the right people for it. If we’re not, we’ll say so in the first reply rather than the third meeting.
Emailinfo@integrateconsulting.comDelivered wherever you operate
Typical response inside one business day