History may not repeat itself, but it rhymes, according to Mark Twain reputedly.

Earlier this week, I experienced one of those rhymes when Finastra CEO Simon Paris cited an unnamed CEO of a large Asian bank as saying “We have to be a software company that does banking.”

It was the same mantra a decade and a half ago for many banks, and which did not end well.

During boom times when revenue is flowing in hand over fist, there is far more tolerance to develop and build systems internally. However, when the markets sour, everything changes: Headcounts shrink, capital spending turns into operational spending, and firms put pet projects on hold.

The decision to build or buy innovative technologies has come around once again with the surge in fintech startups. Banks are feeling competition on almost every business from the approximately 8,000 fintech startups in the US alone.

Many firms’ instincts are to compete with these startups directly and set up their centers of innovations and appoint chief innovation officers, but those strategies seldom prove as fruitful as first expected.

Much of the time, internal innovation teams come up with technologies in search of a problem rather than addressing problems with existing or novel technologies. Can you say “blockchain?”

Another challenge of internally developing artificial intelligence-, blockchain-, machine-learning offerings is the shortage of resources and expertise.

How can one firm’s innovation team compete with the aggregated levels of innovation being brought to bear by the fintech startup community with its “fail fast” development models?

Trying to compete with them is a waste of time and resources. Firms would do better if they let the startups fail or succeed and then deal with the top performers. It provides the greatest bang for the fintech buck and lets banks focus on their core competency, which being banks.

Those who learned this during the first Internet boom have moved on, and it is up to the current generation not to make the same mistakes.