9 Integration with Oracle Digital Assistant
LPO is integrated with a voice assistant such as Oracle Digital Assistant to handle complex workflow and business process questions. The functionality integrates voice assistants, machine learning techniques such as NLP, and a pricing solution such as LPO. This reduces the workflow for a typical user of such software and helps in handling complex business process use cases. For example, buyers or managers who manage price recommendations for retail apparel might be interested in the following use cases:
Sample Conversation: Chain/Country Level View
For example, a user can ask to see the performance of a department or all departments at the country level.
For All Departments
User: "I would like to see how my stores in US are performing."
Bot: "Okay. It seems as if you want to take a look at the country-level metrics for stores.
Please specify the location from the following: <America, Europe, E-commerce US>?"
User: "America."
Bot: "Do you want to review a specific department or all departments?"
User: "All departments."
Bot: "Aggregated metrics for the specified location(s) and all departments are: Revenue is
2756451.736, Change in Revenue is -1.49%, Gross Margin is 1868901.003, Change in Gross Margin is -3% , Sell Through is 0.6861, Change in Sell Through is 1.28%."
For Selected Departments
User: "I would like to see how my stores in US are performing."
Bot: "Okay. It seems as if you want to take a look at the country-level metrics for stores.
Please specify the location from the following: <America, Europe, E-commerce US>?"
User: "America."
Bot: "Do you want to review a specific department or all departments?" User: "Specific Department."
Bot: "Which department are you interested in?"
User: "Womens Activewear"
Bot: "Aggregated metrics for the specified location(s) and department(s) are: Revenue is 2756451.736, Change in Revenue is -1.49%, Gross Margin is 1868901.003, Change in Gross Margin is -3% , Sell Through is 0.6861, Change in Sell Through is 1.28%."
Sample Conversation: Exception Management
For example, a user can ask to see the locations for a department where the margin improvements are lower than 5%.
User: "I want to see the locations that do not meet the approval criteria."
Bot: "Okay. It seems as if you are interested in finding exceptions. First, which department are you interested in?"
User: "Womens Activewear."
Bot: "What metric are you looking for: Revenue Improvement, Margin Improvement, Sell-through Improvement, or Sell-through?"
User: "Sell-through."
Bot: "Please specify the threshold (for example, 15%) for identifying the poorly performing locations."
User: "10%."
Bot: "The location for Department Womens Activewear with Sell-through below 10% is Company."
Bot: "Would you like to see the run summary for this Department and the location?"
User: "Yes."
Bot: "The metrics for run Copy_Of_test are: Revenue is 2756451.736, Change in revenue is
-1.49%, Gross Margin is 1868901.003, Change in gross Margin is -3% , Sell through is 0.6861, Change in sell through is 1.28%."
Bot: "Would you like to approve the run - Yes/No?"
Depending on what the user answers here, the flow will be different. If the user says "Yes," then the user is taken to the Approval Use Case. If the user says "No," then the user is taken to Recalculate Use Case.
Sample Conversation: Approval Process
A few use cases are possible. A user can approve all batch runs or select specific runs for approval. For example, a user can reject the price recommendations for runs that have a discount higher than 40% or approve all batch runs that show an improvement of 5% for revenue.
Simple Runs
User: "I want to review the metrics and would like to take care of run approvals today."
Bot: "Okay. It seems as if you are interested in approving runs. Do you want to approve all batch runs or a single location?"
User: "Single location."
Bot: "Which location are you interested in?"
User: "Company."
Bot: "Which department are you interested in?"
User: "Womens Activewear."
Bot: "The metrics for the run Weekly Run 04242017 are: Revenue is 2756451.736, Change in
Revenue is -1.49%, Gross Margin is 1868901.003, Change in Gross Margin is -3% , Sell
Through is 0.6861, Change in Sell Through is 1.28%."
Bot: "Would you like to approve the run - Yes/No?"
User: "Yes."
Bot: "The run has been approved, and the offers have been sent to the customers."
Bot: "Please let me know how can I help you with the LPO application."
All Batch Runs
User: "I want to review the metrics and would like to take care of run approvals today."
Bot: "Okay. It seems as if you are interested in approving runs. Do you want to approve all batch runs or a single location?"
User: "All batch runs."
Bot: "Which department are you interested in?"
Sample Conversation: Recalculate Use Case
User: "Womens Activewear."
Bot: "All five batch runs have been approved."
Bot: "Please let me know how can I help you with the LPO application."
Sample Conversation: Recalculate
Recalculate the key performance metrics. For example, a user can reject recommendations that have a discount higher than 70% and recalculate the metrics.
User: "I want to review the metrics and would like to see whether to make any changes to the recommendations."
Bot: "Okay. It seems as if you are interested in accepting or rejecting recommendations. First, which location are you interested in?"
User: "Company."
Bot: "Which department are you interested in?"
User: "Womens Activewear."
Bot: "The metrics for run Copy_Of_test are Revenue is 2756451.736, Change in revenue is
-1.49%, Gross Margin is 1868901.003, Change in gross Margin is -3% , Sell through is 0.6861, Change in sell through is 1.28%."
Bot: "Would you like to approve the run - Yes/No?"
User: "No."
Bot: "Would you like to ACCEPT/REJECT available recommendations? - ACCEPT/REJECT."
User: "ACCEPT."
Bot: "Would you like to ACCEPT promotion/markdown recommendation? - PROMOTION/ MARKDOWN."
User: "Promotion."
Bot: "What is the maximum% discount criterion to accept Promotion recommendations?" User: "40%."
Bot: "Promotion Price recommendations below 40% have been accepted."
Bot: "Would you like to recalculate the metrics again - Yes/No?"
User: "Yes."
Bot: "The run has been submitted for recalculating the metrics. To see the new metrics, please query the bot."
Bot: "Please let me know how I can help you next with the LPO application."
- Q&A. This integration helps in answering a user's most frequently asked questions. For example, a user can ask about promotions and the Q&A tab brings up the relevant questions and answers related to the term promotions.
A user can enable the Oracle Digital Assistant by clicking the microphone symbol. A user can say a particular question (utterance) to a voice-enabled device, and the utterance is converted into text and mapped to an intent/entity using existing Natural Language Processing algorithms such as Oracle Intelligent Bots. Once the intent/entity has been identified, it is sent to a pricing solution such as LPO, which then performs the appropriate processes and returns information.
Some examples of the conversation for the above use cases are provided below. The user utterance is prefixed with User and the response from ODA is prefixed with Bot.
Sample Conversation: Department or Zone-Level Performance Analysis with Explainability
User: What is the performance at loc 1 for 310190_blue?
Bot: The performance at loc 1 for 310190_blue is a revenue of $30,432,810.86, which is higher than the optimal revenue of $29,780,414.02. Explain Show SQL
User: Explain
Bot: The current revenue of $30,432,810.86 for 310190_blue is 1.8% higher than the optimal revenue of $29,780,414.02. This suggests that the product is performing better than expected.
However, without historical data, it's difficult to determine if this is a trend or an anomaly. Nevertheless, the positive difference between current and optimal revenue indicates a strong performance at loc 1 for 310190_blue.
User: Give me optimal revenue for 310200_black for each price zone?
Bot: The optimal revenue for 310200_black varies across different price zones. The highest optimal revenue is $5,862,708.03 in price zone 24, followed by $78,640.64 in price zone 16, $52,362.44 in price zone 19, $25,502.90 in price zone 20, and $18,018.18 in price zone 22. User: Can you give me from highest to lowest?
Bot: The optimal revenue from highest to lowest is $5,862,708.03 in price zone 24, $78,640.64 in price zone 16, $52,362.44 in price zone 19, $25,502.90 in price zone 20, and $18,018.18 in price zone 22.
User: Can you aggregate it over all price zones?
Bot: The aggregated optimal revenue across all price zones is $6,037,232.19.
User: Give me the number of promotions used for 627688040_blue
Bot: Please provide either price zone or location to proceed.
User: Price zone 24
Bot: The number of promotions used for 627688040_blue at price zone 24 is 10523.788. Explain Show SQL
User: Explain
Bot: The total promotions used for 627688040_blue at price zone 24 is 10523.788, indicating a significant promotional activity for this product at this price zone. This high number of promotions could be a strategic move to drive sales, clear inventory, or increase brand visibility. However, without prior data, it's difficult to determine if this is an optimal performance or if there's room for improvement. Further analysis on sales data and customer response to these promotions would provide more insights on the effectiveness of this strategy.
User: How about markdowns?
Bot: The total markdowns used is $17,188,948.53.
User: What departments are present in this division?
Bot: The departments present in this division are Women's Activewear, Women's Tops, Women's Bottoms, Beverages, Men's Activewear, Kitchen Appliances, and Men's Tops.
User: What is the optimal revenue for Beverages department?
Bot: The optimal revenue for the Beverages department is $5,349,576.20.
User: Give me the remaining inventory for item 310200_black in all price zones except ones starting with s and i?
Bot: The total remaining inventory for item 310200_black in all price zones except ones starting with s and i is 71091 units.
User: Give me the remaining inventory for item 310200_black for price zone 19?
Bot: The remaining inventory for item 310200_black in price zone 19 is 3682.
User: Give me optimal revenue for 310200_black across all price zones?
Bot: The optimal revenue for 310200_black across all price zones is $6,037,232.19.
User: Explain
Bot: The optimal revenue for 310200_black is $6,037,232.19, indicating a potential maximum earnings for this product. This value can serve as a benchmark for evaluating the product's current performance. To further optimize revenue, it may be beneficial to analyze pricing strategies across different zones and adjust them accordingly to maximize earnings.
Additionally, monitoring sales trends and adjusting inventory levels can help to minimize losses and ensure the product is meeting its full revenue potential. Show SQL
User: Show SQL
Bot: Please give me a few moments to generate your query. Query Generated
User: What is the performance at loc 1 for 310190_blue?
Bot: The performance at loc 1 for 310190_blue is a revenue of $30,432,810.86, which is higher than the optimal revenue of $29,780,414.02. Explain Show SQL
User: Explain
Bot: The current revenue of $30,432,810.86 for 310190_blue is 1.8% higher than the optimal revenue of $29,780,414.02. This suggests that the product is performing better than expected. However, without historical data, it's difficult to determine if this is a trend or an anomaly.
Nevertheless, the positive difference between current and optimal revenue indicates a strong performance at loc 1 for 310190_blue.
User: Give me optimal revenue for 310200_black for each price zone?
Bot: The optimal revenue for 310200_black varies across different price zones. The highest optimal revenue is $5,862,708.03 in price zone 24, followed by $78,640.64 in price zone 16, $52,362.44 in price zone 19, $25,502.90 in price zone 20, and $18,018.18 in price zone 22. User: Can you give me from highest to lowest?
Bot: The optimal revenue from highest to lowest is $5,862,708.03 in price zone 24, $78,640.64 in price zone 16, $52,362.44 in price zone 19, $25,502.90 in price zone 20, and $18,018.18 in price zone 22.
User: What is the optimal revenue for Beverages department?
Bot: The optimal revenue for the Beverages department is $5,349,576.20.
User: Give me the remaining inventory for item 310200_black in all price zones except ones starting with s and i?
Bot: The total remaining inventory for item 310200_black in all price zones except ones starting with s and i is 71091 units.
User: Can you aggregate it over all price zones?
Bot: The aggregated optimal revenue across all price zones is $6,037,232.19.
User: Give me the number of promotions used for 627688040_blue
Bot: Please provide either price zone or location to proceed.
User: Price zone 24
Bot: The number of promotions used for 627688040_blue at price zone 24 is 10523.788.
User: Explain
Bot: The total promotions used for 627688040_blue at price zone 24 is 10523.788, indicating a significant promotional activity for this product at this price zone. This high number of promotions could be a strategic move to drive sales, clear inventory, or increase brand visibility. However, without prior data, it's difficult to determine if this is an optimal performance or if there's room for improvement. Further analysis on sales data and customer response to these promotions would provide more insights on the effectiveness of this strategy.
User: How about markdowns?
Bot: The total markdowns used is $17,188,948.53.
User: What departments are present in this division?
Bot: The departments present in this division are Women's Activewear, Women's Tops, Women's Bottoms, Beverages, Men's Activewear, Kitchen Appliances, and Men's Tops.
User: Give me the remaining inventory for item 310200_black for price zone 19?
Bot: The remaining inventory for item 310200_black in price zone 19 is 3682.
Other Sample Conversations
User: What is life cycle optimization?
Bot: Lifecycle optimization, also known as Lifecycle Pricing Optimization (LPO), is used to determine the optimal pricing recommendations for promotions, markdowns, or targeted offers. It provides answers to questions such as which items to promote, when to promote, how deep to promote, and who to target. LPO caters to retailers interested in optimizing their pricing strategies to maximize profits.
User: How can I delete a run?
Bot: Runs can be deleted, depending on the status of the run and the permissions a specific user has. A user with Administrator permissions can delete any run except for one that has a status of Running. Users can delete any runs they have created that do not have a status of Running. They cannot delete runs created by another user or runs with an Auto run status. More than one run can be selected for deletion at the same time.
User: where is the task pane located?
Bot: The Task Pane is located on the left side of the application. It is used to navigate between LPO Overview, Forecasting, and Manage Recommendations. Users can also use this to navigate between the applications.
User: why am I getting a run not optimized error?
Bot: You are getting a "Run Not Optimized" error because the optimization is not complete. This error occurs when the run is in the "Worksheet" status, and you try to go to the "Results" stage. The optimization process is not finished, and therefore, the results are not available. Additionally, the error can also occur if there are mapping errors in the assortment data, or if the POG set has multiple assortments mapped to it and at least one of them has mapping errors.
User: what is pro_merch_rpm_flg, whats the default value?
Bot: PRO_MERCH_RPM_FLG is a configuration parameter that serves as a flag to identify
whether to publish markdown recommendations to RPM. The default value of PRO_MERCH_RPM_FLG is N.
User: what are the configurations for integrating with pricing cloud services
Bot: To integrate with Pricing Cloud services, additional configuration parameters must be set. These parameters include the Pricing CS username and password, which must be added under the credential store for Merchandise Pricing CS Access. The credential store can be managed using the UI screen for Control and Tactical Center -> Manage Credential Stores.
The configuration parameters for integration with Retail Pricing Cloud Service are listed in Table 12-5, which includes parameters such as IO_MERCH_RMS_REPL_PATH, IO_MERCH_RMS_VERSION, and IO_MERCH_RMS_AUTH_SERVICE_TYPE. These parameters are used to determine the path of the Rest service, the RMS version to use for publishing replenishment parameters, and the authentication type to use for IPO-IO RMS integration.
User: what are the configuration parameters for integrating with pricing cloud services
Bot: The configuration parameters for integrating with Pricing Cloud services are listed in Table
12-5. These parameters include PRO_MERCH_RPM_FLG, PRO_MERCH_RPM_HOSTNAME, PRO_MERCH_RPM_CONTEXT_ROOT, PRO_MERCH_RPM_CLEARANCE_PATH, PRO_MERCH_RPM_PORT, PRO_MERCH_RPM_HTTP_PROXY_HOSTNAME, PRO_MERCH_RPM_HTTP_PROXY_PORT, and PRO_MERCH_RPM_HTTPS_PROXY_PORT. Additionally, the Pricing CS username and password must be added under the credential store for Merchandise Pricing CS Access.
User: what exceptions does LPO Forecasting support?
Bot: LPO forecasting supports exceptions based on sell thru and returns. This includes the sell thru exception, which occurs when the Sell thru end of life is lower than the sell thru exception threshold for any product, and the returns exception, which occurs when the Forecast return % is greater than the return exception threshold for any product.
User: what are the Foundation jobs in POM
Bot: The Foundation jobs in POM are not explicitly listed in your data. However, it does mention that batch jobs are necessary to put the system on a batch schedule and that some of the jobs require relevant configuration parameters to be specified with client-specific values.
User: what are the Foundation jobs in POM that load product hierarchy
Bot: The foundation jobs in POM that load product hierarchy are:
- ORASE_START_BATCH_JOB
- ORASE_START_BATCH_REFRESH_RESTR_JOB
- ORASE_START_BATCH_SET_ACTIVE_JOB
- ORASE_START_BATCH_END_JOB
- RSE_WEEKLY_INPUT_FILES_START_JOB
- WEEKLY_INPUT_FILES_WAIT_JOB
- WEEKLY_INPUT_FILES_VAL_JOB
- WEEKLY_INPUT_FILES_COPY_JOB
- RSE_WEEKLY_INPUT_FILES_END_JOB
- RSE_DIMENSION_LOAD_START_END_JOB
- RSE_DIMENSION_LOAD_START_START_JOB
- RSE_REGULAR_MAIN_LOAD_JOB
- RSE_LOC_SRC_XREF_LOAD_JOB
- RSE_PROD_SRC_XREF_LOAD_JOB
- RSE_LOC_HIER_LOAD_JOB
- RSE_LOC_HIER_TC_LOAD_JOB
- RSE_LOC_HIER_DH_LOAD_JOB
- RSE_FISCAL_MAIN_LOAD_JOB
- RSE_PROD_HIER_LOAD_JOB
- RSE_PROD_TC_LOAD_JOB
- RSE_PROD_DH_LOAD_JOB
- RSE_PROD_GROUP_LOAD_JOB
- RSE_PRICE_ZONE_LOAD_START_JOB
User: How can I allocate budgets for markdowns and promotional activities in our retail strategy?
Bot: To allocate budgets for markdowns and promotional activities in a retail strategy, the system allows for separate budgets for promotions and markdowns. A budget can be fed in at the run's merchandise setup level or it can be fed in at the run's merchandise processing-level. The interface allows the user to provide separate budgets or a combined budget for promotions and markdowns. The relevant fields from the interface are: slspr_rtl_amt (promotion budget), slscl_rtl_amt (clearance budget), loc_exchange_rate, mcal_wid, delete_flg, prod_dh_wid, org_dh_wid.