Due diligence on an AI-enabled BPO business

Due diligence on an AI-enabled BPO business has to answer three questions the seller’s marketing will not settle on its own: does the automation rate actually hold up against real operating data, does every client-required certification survive the specific change of control being proposed, and does the workforce transfer the way the deal structure assumes it will in the province — or country — where staff actually work. These three findings account for most of the deals in this category that either fall apart or get materially repriced after an accepted offer.

The document list that actually matters

• Every client managed-service and master service agreement, along with its renewal history, since a contract renewed several times is a far stronger signal than a favourable term length on paper
• The operating data behind the stated automation rate — transaction logs, ticket volumes, whatever record actually exists — not the summary figure alone
• Employment agreements for the workforce and any documentation of successor-employer obligations that would follow a change of ownership
• Audit records for every client-required security or compliance certification, including whether each one explicitly addresses what happens on a change of control
• Contracts with any third-party AI vendor the automation layer actually depends on, and what happens to pricing or access if that vendor relationship changes

Verifying the automation claim

The most reliable check on an automation-rate claim is comparing it against the underlying transaction or ticket logs directly, rather than accepting a summary percentage prepared for the sale process, and looking at how that rate has moved over several periods rather than at a single snapshot. It is also worth reading the AI vendor contracts closely to understand how much of the automation is actually built and owned by the company versus licensed wholesale from an outside provider, because a business that has built almost nothing of its own is a different acquisition — and carries a different risk — than one with genuine proprietary tooling.

What a certification gap actually means

Finding that a client-required certification does not automatically survive the proposed change of control is a timing problem more than a walk-away problem, but it needs to be treated seriously, because several client contracts are conditioned on the vendor holding that certification and a lapse can put revenue at risk during the re-audit window. The practical response is usually to build the re-audit timeline into the closing schedule, or to hold back part of the purchase price until the re-certification is confirmed, rather than to assume it will simply happen on its own after closing.

What an unsubstantiated automation claim actually means

Finding that the automation rate presented during negotiations cannot actually be reconstructed from the company’s own records is a different kind of problem than finding a modest, honest number — it suggests the figure was never rigorously measured in the first place, and a buyer has no reliable way to know what the real baseline is. In practice this finding usually leads to one of two outcomes: a substantial price reduction that treats the business as closer to a conventional staffing operation until proven otherwise, or a walk-away, if the gap between the marketed story and the verifiable operating reality is wide enough to call the rest of the seller’s disclosures into question.

Employment and workforce jurisdiction checks

Employment-standards and, where relevant, labour-relations legislation governs the workforce in whichever province — or country — it is actually based, and a buyer needs to confirm that jurisdiction rather than assume it matches the company’s head office. Most outsourcing operations are governed provincially rather than federally, though a workforce genuinely embedded in a federally regulated client’s own operations can raise a narrower federal question worth checking rather than assuming away. Where staff are based in Quebec, the province’s labour standards legislation applies rather than any other province’s rules, and a buyer comparing this target against one based in Ontario should not assume the employment obligations transfer the same way. Where any part of the workforce is unionized, confirming whether the collective agreement survives the proposed deal structure is its own separate check, and the answer differs meaningfully between an asset purchase and a share purchase.

Reading a workforce jurisdiction risk finding

A workforce concentrated in a jurisdiction facing rising minimum-wage or labour-standard changes is not automatically a reason to walk away, but it is a finding that belongs explicitly in the disclosure schedule and in the buyer’s own forward projections, not left as an assumption baked into the trailing financials. The practical question is whether the pricing already reflects where labour costs are heading, or whether it was built on a cost base that is about to move — and the disclosure schedule is where that gap gets captured and, often, priced.