Carreker Corporation Press Release - June 3, 1997

New System Frees Non-Earning Assets for Lending and Investments

(DALLAS, TX June 3, 1997) Carreker-Antinori, Inc. announced today the release of its CashForecasterTM system. CashForecaster uses neural-net technology developed by Knowledge Based Systems, Inc. (KBSI), College Station, Texas, to accurately forecast physical cash requirements each day. This technology facilitates a significant and permanent reduction in cash requirements. Carreker-Antinori is a financial services technology firm, providing America's largest banks with end-to-end solutions through synergistic software, consulting and payment operations services.

The premise behind CashForecaster is that the accurate prediction of future cash needs will enable a bank to minimize the amount of cash on hand in vaults, branches and ATMs, thereby eliminating the funding cost of this non-earning asset. "If a bank has a branch, it probably has excess cash in the system, and that cash can be turned into an earning asset," said John Philliben, deposit processing operations manager, Michigan National Bank. "We expect to convert $30 million in non-earning assets into $1.5 to $2 million in additional income to our bottom line."

CashForecaster enables a bank to predict and monitor its inventory of cash by denomination mix. This allows branches to significantly reduce the cushion of excess cash needed to avoid running out of certain bills. With this new technology, the management of branch cash can be moved from the branch to the vault, which determines the amount of cash distributed to the branches daily. "The vault can anticipate changes occurring in demand levels at the branch level even before the branch manager does," Philliben said.

CashForecaster determines withdrawal, deposit and transfer patterns over a user-specified period to predict the minimum amount of cash required in various locations at specific times. "The projections are far more accurate than is possible by simple educated guesswork based on long experience at the branch," said Wyn Lewis, managing director, Carreker- Antinori. "The neural-net technology used with the CashForecaster is a rolling projection of what will happen in the future. The system takes the projections and compares them to the actual dollar denominations and amounts used, and if there is an overage or shortage, it automatically adjusts future projections. CashForecaster is a dynamic system that continues to learn from daily experiences."

"Branch managers used to spend a significant amount of time on forecasting our cash requirements, often overestimating or underestimating our needs," Philliben added. "By having less cash on hand, we have less exposure if a robbery were to occur, and by knowing the cash requirements, we eliminate the cost and time spent ordering extra cash. Branch managers can now concentrate on serving customers more efficiently and building business relationships."

CashForecaster is one of four modules in the Carreker-Antinori Liquidity Management suite of products. Other modules include ReserveLinkTM , designed to sweep a high percentage of NOW account balances into non-reservable accounts; ReserveLinkPlus™ , which extends the sweep process from static NOW accounts to the far more volatile other retail DDA and corporate accounts; and LiquidityManager™ , which will optimize the benefit of the other three modules by identifying and reporting new opportunities as they arise.

KBSI develops creative solutions for a wide range of critical problems faced by government, academic, industrial and other organizations. The company partnered with Carreker-Antinori to develop CashForecaster’s neural network technology. KBSI employs advanced information management technologies to help organizations pursuing business process reengineering, electronic commerce, workflow, simulation based operations scheduling, and activity based costing solutions.

For additional information contact Public Relations, Carreker-Antinori, at (800) 486-1981, or fax (972) 661-9301.