PASDACS Team6 min read
Every firm has a version of this story. A partner sees a demo at a conference. The discount is generous if the firm signs this quarter. A year later the licence renews for a system the team uses at a fraction of its capability, the data never fully migrated, and the spreadsheets are back, “just in case”. The software gets blamed. It is rarely the software's fault.
Most regret is locked in before the purchase, in three decisions the firm never consciously made: what problem the system was for, who would own its adoption, and how it would fit the tools already in use.
A demo answers the vendor's favourite question: what can this product do? The firm's question is different: what does our work need? Write that down first, from the people who run the work - which service lines, which volumes, which hand-offs, which reports. Ten specific requirements beat a hundred feature checkboxes, and they turn every demo from a performance into a test.
Two requirements deserve special weight in a CA practice. First, fit with what stays: Tally, the tax software and the document store are usually not moving, so the new system must exchange data with them rather than demand re-keying. Second, the review model: if the product cannot represent maker-checker gates and review queues the way the firm actually works, no amount of configuration will save it.
The licence price is the visible fraction. The honest comparison adds implementation, data migration, training, integrations, the annual increases, and the cost of leaving - what it takes to get your data out in a usable form. A cheaper licence with an expensive exit is not cheaper. Ask every shortlisted vendor the exit question directly; the quality of the answer is itself a signal.
More implementations die in data migration than in configuration. Insist on a migration plan with reconciliation: the new system's balances, client lists and open items signed off against the old one before cutover. Run one full cycle in parallel where the workflow is critical. And keep the old system readable until the second filing season has passed without a look back.
Software adopted by memo fails quietly. Somebody inside the firm - named, resourced and interested - owns the rollout: training on the firm's own engagements rather than demo data, the first month's questions, and the honest usage numbers afterwards. If nobody can be named, the firm is not ready to buy, whatever the discount says.
One structural fix removes most of the trap: separate the advice from the sale. A recommendation from someone holding a reseller quota is a sales channel; a recommendation from someone who profits only from the fit is advice. Whoever the firm asks, the test is the same - would this adviser earn anything if you chose the product they recommend? If the answer is yes, get a second opinion from someone whose answer is no.
The pattern across all of it: the firm that writes its requirements, prices the exit, plans the migration and names the owner has made regret structurally difficult - before a single demo is booked.
Keep reading
The question is not whether AI enters the practice - staff are already pasting client material into consumer tools. The question is whether it happens under the firm's rules or around them.
22 July 2026 · 7 min read
Past a certain size, the constraint is no longer effort - it is that GST, audit and accounting each run their own way, and every decision routes through a partner. The fix is an operating model, not overtime.
15 July 2026 · 7 min read
Reconciliation absorbs experienced hours every single month, but a CA firm cannot hand tax positions to a machine. The answer is a control boundary, drawn before the first rule is written.
8 July 2026 · 7 min read
If this piece raises questions about your own operations, our team is happy to discuss what it could mean in practice.