Core Banking Implementations Shorten With AI Agents
Summary
Temenos and Fiserv say AI agents now handle core banking installs, migrations and upgrades; Fiserv cut one product's implementation time by 50%.
Temenos AG (TMSNY) and Fiserv, Inc. (FISV) both said on their third-quarter 2026 earnings calls, held between October 27 and November 3, 2026, that they now use AI agents to install, migrate and upgrade bank core systems in order to shorten core banking implementations[1][2].
Implementation, not software, is the costly part of a core banking switch
A core system is the underlying software in which a bank records deposit and loan accounts and every transaction. When a bank replaces or upgrades it, years of account data have to be mapped field by field to the new system, converted and tested again and again. That work is called implementation, and it usually takes large teams of consultants and engineers and runs for years. Temenos CEO Takis Spiliopoulos said that for many banks the biggest hurdle to starting a project is the cost of implementation rather than the cost of the software[1]. AI agents are AI programs that carry out multi-step tasks toward a goal on their own, and on a core project they can take over repetitive mapping, conversion and testing work. Spiliopoulos argued that shorter projects mean banks earn back their investment sooner, which should help more banks start projects[1]. Every core vendor faces the same implementation hurdle, so similar tools may spread across the industry.
Both vendors moved from trials to agents across several core products
Temenos's position moved noticeably in six months. On its first-quarter call in April 2026, management said it was still early to talk about whether AI could speed up implementations and upgrades[3]. By October the company said it had built many agents itself and with partners, and was running a number of proofs of concept that use AI to install new software and to upgrade existing clients to newer releases, with numbers that were "already quite exciting"[1]. Fiserv has also widened its coverage. On its second-quarter call in August, the company disclosed that implementation timelines for its Cash Flow Central product had been cut by 50%[4]. In November, CEO Takis Georgakopoulos said Fiserv now has multiple agents across key offerings including XD, Cash Flow Central, DNA and Core Advance to dramatically shorten implementation and migration timelines[2]. Neither company disclosed how many days or dollars a single core project saves, and Temenos acknowledged that the tooling is not yet industrial grade[1].
Fiserv's client figures from the same call should be read separately. Measured by revenue, client attrition is down 17% so far this year and client wins are up more than 50%, though Georgakopoulos said the base for wins is low[2]. Management credited the improvement in attrition to service fixes and did not link it to agents, so these numbers are not yet evidence that the agents are working[2].
Lower implementation barriers bring both new orders and client movement
Faster migrations affect both vendor orders and the stickiness of existing clients. For software vendors, more banks may start core replacements and upgrades, which would lift subscription bookings. For vendors with large installed bases, leaving also becomes easier, so share could move faster toward whichever vendor has better service and tooling. Peers do not agree on how much switching gets easier. FIS argues that switching costs remain high because changing a core drags debit, card, print and mail services and state regulation along with it[5]. Jack Henry still quotes 15 to 24 months for new clients to go live, says contract terms and staff retraining take the longest, and has added human conversion teams[6]. Agents may shorten the technical work without speeding up contracts or people. The first things to track are the conversion timelines, client attrition and deconversion fees each vendor reports, followed by subscription bookings in 2027 and 2028.
Companies exposed to faster core migrations
- Fidelity National Information Services (FIS): It provides core processing to US regional and community banks and faces the same implementation and conversion workload as Fiserv and Temenos; if migration timelines shorten across the industry, both its client wins and its client losses could speed up.
- Jack Henry (JKHY): It won 58 competitive core deals in fiscal 2026 and collects deconversion fees when clients leave, which it guides to $23 million for fiscal 2027[6]; faster migrations could change both how quickly its new wins go live and how firmly it holds its own clients.
- Fujitsu (6702.T): It builds and migrates core systems for Japanese banks, and its modernization orders rose 56% to JPY239.9 billion in the first half of fiscal 2026[7]; shorter core projects could change the staffing and pricing of this kind of work.
Sources
[1] Drillr · Temenos AG · 2026-10-27 · Q3 2026 earnings call
"Because it is still one of the biggest hurdles for many banks to embark, yeah, is the cost of implementation, not so much the cost of the software."
[2] Drillr · Fiserv, Inc. · 2026-11-03 · Q3 2026 earnings call
[3] Drillr · Temenos AG · 2026-04-21 · Q1 2026 earnings call
[4] Drillr · Fiserv, Inc. · 2026-08-06 · Q2 2026 earnings call summary
[5] Drillr · Fidelity National Information Services (FIS) · 2026-05-08 · Q1 2026 earnings call
[6] Drillr · Jack Henry (JKHY) · 2026-08-19 · Fiscal 2026 Q4 earnings call
[7] Drillr · Fujitsu (6702.T) · 2026-10-29 · Fiscal 2026 Q2 earnings call
This article only helps surface industry changes and companies that may be overlooked. It is not a stock recommendation.