Eltropy’s war on phone tag

At Centric Federal Credit Union, a loan application could take 48 hours. The culprit in Eltropy’s account of the case was familiar: unanswered emails, phone tag, paperwork waiting for someone’s attention. Then Centric put secure text links and DocuSign into the conversation. Reported application time fell to one hour. Some applications took ten minutes. The striking detail is how little science fiction this required. Someone needed a signature. Someone else needed a convenient way to provide it.

That is a useful place to begin with Eltropy, because its expanding catalog can obscure its original appeal. Banking contains a remarkable amount of waiting disguised as work. A member misses a call. A loan officer leaves a message. An application sits unfinished. Better communication can change the economics before anyone changes the financial product.

Read the Centric result carefully. It describes an application process, with an unusually fast subset alongside the broader improvement. It does not say that every borrower receives a decision in ten minutes. The interesting question for another lender is which part of its own elapsed time belongs to actual evaluation and which part belongs to waiting for a reply. A faster message cannot rewrite underwriting policy. It can make the next required step easier to finish while the member is still paying attention.

01 Two founders, one ordinary problem

Ashish Garg and Chandra Bhople founded Eltropy in 2013. Both attended BITS Pilani, although 23 years separated their time there. Garg’s father had been a bank clerk. In a 2021 interview, Garg described leaving a semiconductor career after completing an MBA and obtaining his green card. His wife, Ranjeeta, took on the household’s financial responsibilities while he pursued the company.

There is a pleasingly academic joke in its name. The founders took “entropy” and substituted an L for learning. Their explanation connected learning with the orderly transfer of knowledge. Today, that idea has a practical banking interpretation: get the right information to the person who can use it.

02 Buying the missing pieces

Eltropy grew beyond the text message by buying capabilities that sit next to it. In June 2022, it acquired POPi/o, a video banking company. Video made it possible to put a remote employee in front of a member and work through documents together. In August, it bought Marsview.ai, adding conversational intelligence and automation. In January 2025 came Lexop, a collections technology provider.

The sequence matters. A reminder about a late payment has limited value if paying remains awkward. Lexop brought self-service repayment technology closer to Eltropy’s communication channels. The company was extending its reach from getting someone’s attention to helping them act. For institutions managing separate communication vendors, the commercial attraction is fewer disconnected tools and handoffs.

03 An answer is only half the job

The current platform combines secure messaging, live chat, voice, video, co-browsing, and contact center tools. Its AI Assistants help employees retrieve answers from institutional knowledge. AI Voice and AI Chat handle routine member interactions. The distinction is useful: an employee looking up a policy and a member checking an account need different kinds of assistance.

In March 2026, Eltropy announced its Agentic AI Platform. The ambition is to carry a request into banking workflows through connected systems. Constant AI, for example, is building a Skip-A-Pay agent on the platform. Eltropy’s proposed controls include authorized access, activity logs, data boundaries, and limits set by approved operating procedures. Those controls are product commitments that institutions must evaluate in practice.

A September 2026 Tyfone integration shows the same thinking in a smaller package. A customer already logged into digital banking can enter Eltropy support without repeating authentication. Chat, video, and co-browsing stay inside the banking experience. The software carries forward something the institution has already established: who this person is.

“A member shouldn’t have to prove who they are twice.”

Abhishek Tiwari, Eltropy chief product officer

04 The price of finishing

Eltropy sells business software to institutions. It is not a bank, and members do not buy its platform as a consumer subscription. Its June 2021 financing announcement described a $25 million Series A led by K1 Investment Management, with Curql Fund and CMFG Ventures participating. Credit union investment funds were helping finance a supplier to their own industry.

That relationship gives Eltropy a particular place in the market, but plenty of company. Glia combines banking AI with voice and digital service. Posh offers AI for member interactions and employees. Banking specialization alone will not settle the comparison. Eltropy’s case rests on how its communication channels, integrations, and operational tools work together for a particular institution.

The sensible buying exercise is a complete implementation quote: software, integration work, staff training, and the cost of running it. Compare that total with the old process. A shorter call means little if an employee must spend the saved minutes repairing the transaction afterward. Count completed requests, repeat contacts, and total staff time.

A useful pilot would keep the test modest. Choose one recurring request, agree on what counts as completion, and record where employees still intervene. Check the member experience as well as the dashboard: did someone have to repeat a story, find another document, or call back? A high automation rate can conceal an inconvenient service. For a community institution, the relationship is part of the product, so an apparently efficient process deserves scrutiny if members find it harder to get help.

05 Copy the handoff, then count

There is a lesson here that requires no software purchase. Take one unfinished request and follow it. Find the first avoidable delay. Centric’s example points to the space between reaching a member and receiving a signed document. Tyfone’s points to authentication repeated after login. Both suggest that service improves when useful information survives the handoff.

The conditions matter. Automation needs current policy knowledge, access to the relevant systems, and a clear route to a person. An institution with few repetitive requests may struggle to justify a broad deployment. A workflow that cannot safely reach its system of record may produce a fluent conversation and leave the work untouched.

Eltropy’s proposition is easy to test, at least in principle. Pick a real banking request. Watch it travel. Then ask whether the member got what they came for, and how much effort everyone spent getting there.