A field note for business owners tired of being told blockchain will fix everything
In 2018, Maersk and IBM announced TradeLens — a blockchain platform designed to digitise global shipping documentation and finally eliminate the mountains of paper that slow down every container crossing an ocean. Dozens of partners signed up. Every conference deck for two years used it as “blockchain, finally real.”
By 2023, it was shut down.
Not because the technology failed. Because competing shipping lines didn’t want to hand their data to a platform co-owned by one of their own rivals. The tech worked fine. The trust didn’t.
I open with that story because it’s a useful corrective. Every few years, someone predicts blockchain will “revolutionise contracts,” and five years later the revolution has quietly become a pilot project, then a case study, then a slide in someone’s old deck. If you’ve sat through enough of these cycles as a business owner, a healthy scepticism is earned, not cynical.
But something genuinely different is happening this time — and it isn’t blockchain doing it alone. It’s blockchain finally getting a partner that can do the part it was always bad at: understanding messy, real-world, non-binary information. That partner is AI. And it changes the five-year forecast quite a bit.
What blockchain actually got right
Strip away the hype, and blockchain solved one specific problem well: letting parties who don’t fully trust each other agree on a single, tamper-proof version of events, without a middleman adjudicating every step.
That’s precisely why it has worked in narrow, well-defined settings, and HSBC’s trade finance work is a good illustration, not a hypothetical one. In November 2018, HSBC India and ING Bank Brussels ran a live letter-of-credit transaction on R3’s Corda network — ING issuing the credit for US-based importer Tricon Energy, with HSBC India as the advising and negotiating bank for exporter Reliance Industries. Reliance’s own CFO put a number on it afterwards: a document exchange that normally runs seven to ten days, compressed to under a day. HSBC took the same approach further in April 2021, running a live blockchain trade on the Contour platform between Tata Steel (India) as exporter and Universal Tube & Plastic Industries (UAE) as importer, with HSBC UAE issuing the letter of credit and HSBC India again advising for the exporter — a global first for the steel industry, and one HSBC says cut the same 5–10 day document cycle down to under 24 hours. Neither was a proof-of-concept run for a press release. Both were live trades, with real steel and real soybeans-and-tuna-style cargo behind them, real banks on both ends, and named executives willing to put their names to the results.
Notice what’s common to the successes and absent from TradeLens: a small, defined circle of counterparties, a narrow document set, and a clear financial incentive for everyone at the table to cooperate. Blockchain thrives inside those boundaries. It struggles the moment you ask an entire, competitive industry to share a ledger.
The part blockchain was never going to solve — and AI just did
Here’s the limitation nobody likes to admit: a smart contract only executes correctly if the conditions it’s checking are unambiguous. “If payment received, release goods” works beautifully on a blockchain. “If the goods arrived in acceptable condition” does not — because “acceptable condition” isn’t a yes/no fact sitting in a database. Someone, or something, has to interpret it.
That’s the gap AI is now filling, in two distinct ways.
First, on the drafting and negotiation side. Contract review platforms — Harvey, Ironclad, Icertis, Luminance, Spellbook, and DocuSign’s Agreement Cloud (which absorbed Lexion), among others — have moved from “search this PDF for a clause” to reading an entire contract set, flagging where a new vendor agreement conflicts with an existing data processing agreement, and drafting a redline against your playbook. Legal teams report that contract review, which used to consume hours, now takes a fraction of that. This doesn’t touch blockchain at all, but it’s the necessary precondition: you can’t put a contract on-chain until AI (or a very patient lawyer) has made its terms machine-readable in the first place.
Second, and more interestingly, on the verification side. Chainlink — the oracle network that feeds real-world data into blockchain smart contracts — has spent the last two years building infrastructure specifically to let AI models certify the fuzzy, real-world conditions that smart contracts can’t check on their own. Banks are already piloting Chainlink’s AI-assisted oracles to extract and verify data from unstructured documents like PDFs and scanned filings, so that a smart contract can act on something that used to require a human to read and sign off on. That’s the genuinely new piece: AI as the judgement layer, blockchain as the execution and record-keeping layer.
The law has quietly caught up, too
It’s worth knowing that this isn’t sitting in a legal grey zone waiting for lawmakers to catch up. In 2021, the UK Law Commission concluded, after a formal review, that English law already supports smart legal contracts as legally binding agreements — no new legislation required, provided they meet the same basic tests as any contract: agreement, consideration, and an intention to be bound. If your business has any UK-linked contracts, this matters more than it sounds — the legal plumbing is quietly ready even where the business habits aren’t.
What five years out actually looks like
Not “every contract on a blockchain.” That prediction has been wrong for a decade and will still be wrong in 2031. What’s more realistic is a three-tier split:
- Routine, parametric agreements — trade finance documentation, insurance payouts triggered by verifiable events, royalty and licensing payments, vendor SLAs with measurable KPIs — are increasingly self-executing, with AI verifying the trigger condition and blockchain handling settlement and the audit trail.
- Standard commercial contracts — the bulk of what any mid-sized business signs — are largely AI-drafted and AI-reviewed against a playbook, with humans approving rather than redlining line by line, and blockchain is used selectively where multiple parties need a shared, tamper-proof record.
- High-stakes, bespoke deals — M&A, complex partnerships, anything hinging on a negotiated definition of “reasonable” or “material” — stay firmly in human hands. AI will draft the first pass and flag the risk clauses; blockchain, if it appears at all, will just be the record-keeping layer underneath. No amount of automation replaces judgment when the terms themselves are the product of negotiation, not a formula.
None of this is the “smart contracts replace lawyers” story that made the rounds a few years ago, and it isn’t the “blockchain revolutionises everything” story either. It’s narrower and, I’d argue, more useful to actually plan around: blockchain keeps doing the one thing it’s good at — a shared, tamper-proof record — while AI takes on the judgment calls that used to require a person to read, interpret, and sign off. For most businesses, the practical question isn’t “should we get on a blockchain?” It’s “which of our recurring, rule-based agreements are ready to have their trigger conditions verified by AI instead of a person” — because that’s the part of this shift you can actually start planning for now.
As always — verify any figure here against the source before you build a decision on it.
Sources: HSBC/ING press release via Finextra and Ledger Insights (Nov 2018); HSBC/Tata Steel press release via Global Trade Review and HSBC UAE (Apr 2021); UK Law Commission, “Smart Legal Contracts: Advice to Government” (2021).
