- SAP Community
- Groups
- Interest Groups
- Artificial Intelligence and Machine Learning
Group Information
Artificial Intelligence and Machine Learning
Join peers and experts to discuss advances in AI, use cases developed by SAP, and explore the limitless possibilities of AI built for businesses.
1216 members
Open group
Artificial Intelligence and Machine Learning
Join peers and experts to discuss advances in AI, use cases developed by SAP, and explore the limitless possibilities of AI built for businesses.
Group Information
Artificial Intelligence and Machine Learning
Join peers and experts to discuss advances in AI, use cases developed by SAP, and explore the limitless possibilities of AI built for businesses.
1216 members
Open group
Created 2023 Sep 04
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for
Featured Topics
Recent Activity
Product and Topic Expert
2,141
2
11
This blog shows how to implement a SAP Business AI project with a time-series forecast by using the embedded Predictive Analysis Library. Implementation is possible with the free-tier edition of SAP HANA Cloud
... View more
2024 Aug 232:29 PM
Posted in Artificial Intelligence and Machine Learning Blogs
Product and Topic Expert
1,273
4
2
Unleash Open-Source Text Embeddings in SAP AI Core with Infinity This blog post dives into using open-source Text Embedding Models within SAP AI Core. It explores the "Bring Your Own Model" (BYOM) approach, focusing on Infinity – a high-performance server for serving text embeddings generated by various models. Key Takeaways: Integrate open-source Text Embedding Models into SAP AI Core for cost-effective text embedding. Leverage Infinity's model-agnostic nature for a wide range of models. Deploy custom inference servers like Infinity for greater control and flexibility. Benefits of Infinity with SAP AI Core: Scalable: Scales efficiently with increasing demands on your SAP AI Core infrastructure. Transparent: Open-source nature fosters transparency and allows for customization. Ready to Get Started? Explore the possibilities! Refer to the provided resources, including sample code, to deploy and run open-source LLMs with Infinity in SAP AI Core. This blog post covers: A quick introduction to Infinity Deploying and running Infinity within SAP AI Core Inferencing a text embedding model with Infinity Key considerations for using Infinity with SAP AI Core Go deeper! Explore the full blog post for a step-by-step guide on deploying and using Infinity within SAP AI Core to unlock the power of open-source text embeddings for your projects.
... View more
2024 Jul 0212:32 PM
Posted in Artificial Intelligence and Machine Learning Blogs
Product and Topic Expert
222
0
1
In my previous blog post, we have discussed the business values, overview and technical details about end-to-end Low-Code Machine Learning with Ludwig and SAP AI Core. In this blog post, I'll continue sharing my experience of replicating another two custom AI use cases from previous work steam with ludwig in low-code fashion, comparing with their original implementation in pro-code approach.
... View more
Tuesday
Posted in Artificial Intelligence and Machine Learning Blogs
ODATA DRAFT SCENARIOS What is Draft? A draft is the state of the business object when its data is not persisted in the actual tables but is stored in some temporary tables which can be quickly retrieved to continue working (from the place left) in case the work is left in the middle due to some reason. Draft Management Process Flow: Consider an application with Sales Order (SO) Header and Item data. The data can be displayed and updated or new SOs can be created. Drafts play no role in Display functionality. It is when any user either enters into an update mode or a create mode drafts are involved. A Draft can be: Created when the user enters into update mode – all actual table data is copied for temporary tables/draft tables and then data is presented on the screen for the user to update. Created when the user enters into create mode- table entries with draft keys are created in draft tables. Updated when the user enters data in any field, changes a tab on the screen, on any clicks, the press of a button, or at a set time depending on the requirement of the application. Any one of the trigger points or a combination of these trigger points can be used to implement this part. Deleted when a draft is discarded by the user, or when the data is saved to actual tables. Implement Draft Management: Create Draft tables to store the data temporarily. In this case, draft tables will be require for Header and Item data. Tables for draft data: Steps to create Entity types: Go to SAP Gateway Service Builder(SEGW). Click on create project node. Provide the project name, description, package name and click on Continue. Select the Data model and Right click on data model. Select Import and click on DDIC Structures node. Provide the entity name and DDIC table name which table you want to import. Click on Next. Select the required fields which you want to display/perform operations. Click on Next. Select the atleast one key fields from that structure and click on finish. Create one more entity type for Item details. Click on Save. Click on Generate Runtime Objects node. By default/Implicitly 4 classes, one technical model and one technical service is created. Click on continue. In addition to this one more entity with just a text field is defined to be used as the return type Create one return type structure. Add one return type. Steps to create Association : Right click on Data model. Select Create and click on Association node. Provide the Association name, Principle name as Header entity name, Dependent name as Item entity name and provide the cardinality for displaying the item details and provide the navigation property name. Click on Next. Provide dependent property name as common field name for both header and item entity. Click on next. Click on Finish. Create one deep structure in MPC_EXT class. Provide Navigation property name for item details like as shown in below Steps to create Function Import: 3 function imports, CreateDraft, DeleteDraft, and SaveOrder are created to achieve the functionality. All use the Return Type as the entity type Return. Right click on Data model. Select the create and Click on Function import. Provide the Function import name, Click on continue. Provide the return type kind as Entity type, Return type as Return entity name, pass the cardinality and select the HTTP method as post. Like this create 2 more function import for DeleteDraft, and SaveOrder. Provide the parameter for function import. Code Reference from ZCL_ZGR_ODATA_DRAFT_SC_DPC_EXT Read and Query Operations: These operations usually will refer to the actual tables/Standard BAPI etc., to get data. method SO_HEADERSET_GET_ENTITY.
DATA : lv_vbeln TYPE vbak-vbeln.
DATA(it_keys) = io_tech_request_context->get_keys( ).
lv_vbeln = VALUE #( it_keys[ name = 'VBELN' ]-value optional ).
lv_vbeln = |{ lv_vbeln ALPHA = IN }|.
select SINGLE FROM vbak
FIELDS vbeln,
erdat,
erzet,
ernam,
netwr,
waerk,
vkorg,
vtweg,
spart,
kunnr
WHERE vbeln = _vbeln
INTO CORRESPONDING FIELDS OF _entity.
endmethod. method SO_HEADERSET_GET_ENTITYSET.
SELECT FROM vbak
FIELDS vbeln, erdat, erzet, ernam, netwr,
waerk, vkorg, vtweg, spart, kunnr
INTO CORRESPONDING FIELDS OF TABLE _entityset
UP TO 20 ROWS.
endmethod. Read/ Get Item data: method SO_ITEMSET_GET_ENTITY.
DATA ls_keys TYPE zcl_zgr_odata_draft_sc_mpc=>ts_so_item.
io_tech_request_context->get_converted_keys(
IMPORTING
es_key_values = ls_keys " Entity Key Values - converted
).
SELECT SINGLE FROM vbap
FIELDS vbeln, posnr, matnr, arktx, zmeng, meins,
netwr, waerk, werks, lgort, vstel, route
WHERE vbeln = _keys-vbeln
AND posnr = _keys-posnr
INTO CORRESPONDING FIELDS OF _entity.
endmethod. METHOD so_itemset_get_entityset.
DATA lv_vbeln TYPE vbak-vbeln.
IF io_tech_request_context->get_source_entity_type_name( ) = 'SO_Header'.
DATA(It_Keys) = io_tech_request_context->get_source_keys( ).
lv_vbeln = VALUE #( it_keys[ name = 'Vbeln' ]-value OPTIONAL ).
lv_vbeln = |{ lv_vbeln ALPHA = IN }|.
SELECT FROM vbap
FIELDS vbeln, posnr, matnr, arktx, zmeng, meins,
netwr, waerk, werks, lgort, vstel, route
WHERE vbeln = _vbeln
INTO CORRESPONDING FIELDS OF TABLE _entityset.
ELSE.
SELECT FROM vbap
FIELDS vbeln, posnr, matnr, arktx, zmeng, meins,
netwr, waerk, werks, lgort, vstel, route
INTO CORRESPONDING FIELDS OF TABLE _entityset.
ENDIF.
ENDMETHOD. CREATE_DRAFT Method: An additional method is created in the DPC_EXT class with importing parameter VBELN i.e., Order Number and Exporting parameter Return which is the same type as extra entity. This method will check weather the draft exists for the order. If it does and it is by another user, it will return an error text. Otherwise, the method will copy the data from the standard table to custom draft tables ZGR_T_SOH_DRAFT and ZGR_T_SOI_DRAFT. To Keep the logic simple, the method uses incremental logic to generate a key. In an actual scenario, we will either use a hash generator or number range to create the draft key. Note: Create_Draft method is a custom method which is have importing and Returning parameter. method CREATE_DRAFT.
"Check if draft exists
SELECT SINGLE FROM zgr_t_soh_draft
FIELDS *
INTO (ls_draft).
IF sy-subrc EQ 0. "Draft exist
IF sy-uname = ls_draft-ernam. "Draft by same user
RETURN = VALUE zgr_s_return( test = 'S Draft success' ).
ELSE.
RETURN-test = |E Draft user : { ls_draft-ernam }|.
ENDIF.
ELSE. "No Draft
SELECT SINGLE FROM vbak
FIELDS vbeln, erdat, erzet, ernam, netwr,
waerk, vkorg, vtweg, spart, kunnr
WHERE vbeln = @vbeln
INTO (ls_vbak).
IF sy-subrc EQ 0.
ls_draft = CORRESPONDING #( ls_vbak ).
ls_draft-ernam = sy-uname.
INSERT zgr_t_soh_draft FROM ls_draft.
IF sy-subrc EQ 0.
RETURN-test = 'S Draft Success'.
ELSE.
RETURN-test = 'S Draft Failed'.
ENDIF.
ENDIF.
"Add items to the draft
SELECT FROM vbap
FIELDS mandt, vbeln, posnr, matnr, arktx, zmeng, meins,
netwr, waerk, werks, lgort, vstel, route
WHERE vbeln = @vbeln
INTO TABLE (lt_items).
IF sy-subrc EQ 0.
INSERT zgr_t_soi_draft FROM TABLE _items.
ENDIF.
ENDIF.
endmethod. CREATE Operation: Create Operation should be called only for item data. For header data instead of create, created draft should be called. Item create operation, actually only creates an entry in the draft table. method SO_ITEMSET_CREATE_ENTITY.
DATA: ls_Final TYPE zcl_zgr_odata_draft_sc_mpc_ext=>TY_FINAL,
ls_item type zcl_zgr_odata_draft_sc_mpc=>ts_so_item.
io_data_provider->read_entry_data(
IMPORTING
es_data = ls_final
).
ls_item = CORRESPONDING #( ls_final ).
INSERT zgr_t_soi_draft FROM ls_item.
IF Sy-subrc EQ 0.
er_entity = ls_item.
ENDIF.
endmethod. UPDATE Operation: Update operation for header and item entities will update the draft tables. Update Operation For Header Data: method SO_HEADERSET_UPDATE_ENTITY.
DATA: ls_final TYPE zcl_zgr_odata_draft_sc_mpc_ext=>ty_final,
ls_header TYPE zcl_zgr_odata_draft_sc_mpc_ext=>ts_so_header.
io_data_provider->read_entry_data(
IMPORTING
es_data = ls_final
).
ls_header = CORRESPONDING #( ls_final ).
IF sy-subrc Eq 0.
MODIFY zgr_t_soh_draft FROM ls_header.
er_entity = ls_header.
ENDIF.
endmethod. Update operation for Item data: method SO_ITEMSET_UPDATE_ENTITY.
DATA: ls_final TYPE zcl_zgr_odata_draft_sc_mpc_ext=>ty_final,
ls_item TYPE zcl_zgr_odata_draft_sc_mpc_ext=>ts_so_item.
io_data_provider->read_entry_data(
IMPORTING
es_data = ls_final
).
ls_item = CORRESPONDING #( ls_final ).
IF sy-subrc EQ 0.
MODIFY zgr_t_soi_draft FROM ls_item.
ENDIF.
endmethod. DELETE Operation: In this case, a delete operation is not implemented as a Sales Order can not be deleted from the system. Usually, it will be set as all items rejected etc., which can be done in an update operation. Function Import Implementation: Method /IWBEP/IF_MGW_APPL_SRV_RUNTIME~EXECUTE_ACTION is redefined to implement the function imports. CreateDraft: Call the method Create_draft with the Sales order number. DeleteDraft: Delete the draft data from the header and item draft tables. SaveOrder: Check if the order number exists. If it does, it is a change mode, else it is a create mode. Get data from draft tables. Call create/change BAPI. Return the Message with order number. Function Import implementation: METHOD /iwbep/if_mgw_appl_srv_runtime~execute_action.
DATA : lv_vbeln_out TYPE vbeln.
DATA : ls_header_in TYPE bapisdhd1,
lt_items_in TYPE STANDARD TABLE OF bapisditm,
lt_partners TYPE STANDARD TABLE OF bapiparnr,
lt_return TYPE STANDARD TABLE OF bapiret2.
CASE iv_action_name.
WHEN 'CreateDraft'.
TRY.
DATA(lt_pars) = io_tech_request_context->get_parameters( ).
DATA(lv_vbeln) = VALUE vbeln_va( lt_pars[ name = 'SALESORDER' ]-value OPTIONAL ).
lv_vbeln = |{ lv_vbeln ALPHA = IN }|.
DATA(ls_return) = create_draft( vbeln = lv_vbeln ).
CATCH cx_sy_itab_line_not_found.
"Error
ls_return-test = 'E Invalid call'.
ENDTRY.
copy_data_to_ref(
EXPORTING
is_data = ls_return
CHANGING
cr_data = er_data
).
WHEN 'DeleteDraft'.
TRY.
lt_pars = io_tech_request_context->get_parameters( ).
lv_vbeln = VALUE vbeln_va( lt_pars[ name = 'SALESORDER' ]-value ).
lv_vbeln = |{ lv_vbeln ALPHA = IN }|.
DELETE FROM zgr_t_soh_draft WHERE vbeln = _vbeln.
IF sy-subrc EQ 0.
DELETE FROM zgr_t_soi_draft WHERE vbeln = _vbeln.
ls_return-test = 'S draft deleted'.
ENDIF.
CATCH cx_sy_itab_line_not_found.
"Error
ls_return-test = 'E Invalid call'.
ENDTRY.
copy_data_to_ref(
EXPORTING
is_data = ls_return
CHANGING
cr_data = er_data
).
WHEN 'SaveOrder'.
TRY.
lt_pars = io_tech_request_context->get_parameters( ).
lv_vbeln = VALUE vbeln_va( lt_pars[ name = 'SALESORDER' ]-value OPTIONAL ).
lv_vbeln = |{ lv_vbeln ALPHA = IN }|.
select SINGLE vbeln FROM vbak into (lv_vbeln_Test)
WHERE vbeln = _vbeln.
IF sy-subrc NE 0.
CLEAR lv_vbeln_test.
ENDIF.
SELECT SINGLE FROM zgr_t_soh_draft
FIELDS *
WHERE vbeln = _vbeln
INTO (ls_header).
IF sy-subrc NE 0.
ls_return-test = 'E Sales order not found'.
ELSE.
"Fetch draft items
SELECT FROM zgr_t_soi_draft
FIELDS *
WHERE vbeln = _vbeln
INTO TABLE (lt_item).
IF lv_vbeln_test IS NOT INITIAL. "Change mode
DATA : ls_headerx_in TYPE bapisdh1x,
ls_uheader_in TYPE bapisdhd1,
lt_itemsx_in TYPE STANDARD TABLE OF bapisditmx.
lt_items_in = VALUE #( FOR ls_item IN lt_item
( itm_number = ls_item-posnr
short_text = ls_item-arktx
)
).
lt_itemsx_in = VALUE #( FOR ls_item IN lt_item
( itm_number = ls_item-posnr
updateflag = 'U'
short_text = 'X'
)
).
ls_headerx_in-updateflag = 'U'.
lv_vbeln_out = lv_vbeln_test.
CALL FUNCTION 'BAPI_SALESORDER_CHANGE'
EXPORTING
salesdocument = lv_vbeln_test
order_header_inx = ls_headerx_in
TABLES
return = lt_return
ORDER_ITEM_IN = lt_items_in
ORDER_ITEM_INX = lt_itemsx_in
.
ELSE.
ls_header_in-doc_type = 'ZOR1'.
ls_header_in-doc_date = ls_header-erdat.
ls_header_in-created_by = ls_header-ernam.
ls_header_in-currency = ls_header-waerk.
ls_header_in-sales_org = ls_header-vkorg.
ls_header_in-distr_chan = ls_header-vtweg.
ls_header_in-division = ls_header-spart.
lt_partners = VALUE #( ( partn_role = 'AG'
partn_numb = ls_header-kunnr ) ).
lt_items_in = VALUE #( FOR LS_ITEM IN lt_item
( itm_number = ls_item-posnr
material = ls_item-matnr
short_text = ls_item-arktx
target_qty = ls_item-zmeng
target_qu = ls_item-meins
plant = ls_item-werks
store_loc = ls_item-lgort
ship_point = ls_item-vstel
route = ls_item-route
)
).
CALL FUNCTION 'BAPI_SALESORDER_CREATEFROMDAT2'
EXPORTING
order_header_in = ls_header_in
IMPORTING
SALESDOCUMENT = lv_vbeln_out
tables
RETURN = lt_return
ORDER_ITEMS_IN = lt_items_in
order_partners = lt_partners
.
ENDIF.
IF NOT line_exists( lt_return[ type = 'E' ] ).
"Update FM to delete the draft
CALL FUNCTION 'BAPI_TRANSACTION_COMMIT'
* EXPORTING
* WAIT =
* IMPORTING
* RETURN =
.
"Delete Draft
delete FROM zgr_t_soh_draft WHERE vbeln = lv_vbeln.
delete FROM zgr_t_soi_draft WHERE vbeln = lv_vbeln.
ls_return-test = lv_vbeln_out.
ELSE.
ls_return-test = 'E Error in Bapi Call'.
ENDIF.
ENDIF.
CATCH cx_sy_itab_line_not_found.
"Error
ls_return-test = 'E Invalida call'.
ENDTRY.
me->copy_data_to_ref(
EXPORTING
is_data = ls_return
CHANGING
cr_data = er_data
).
ENDCASE.
ENDMETHOD. Steps to Register the Services: Goto SAP Gateway services (/IWFND/MAINT_SERVICES). Click on Add Services node. Provide the System Alias and Technical service name which is generated while generating the OData services. Click on Get services node. Select the services and click on Add Selected Services. Provide the Package name and click on continue. Select the services by using Filter. Click on SAP Gateway Client. Some basic scenarios output:
... View more
a week ago
Posted in Artificial Intelligence and Machine Learning Blogs
Ludwig is a popular open-source low-code framework for building custom AI models like LLMs and other deep neural networks. It simplifies the machine learning process with a low-code approach, enabling non-ML roles(application developers, business users etc) and empowering the ML professionals(ML Engineers, Data Scientists etc) to quickly experiment, build, train, and deploy models. With integration to SAP AI Core, these machine learning operations can be streamlined and managed at scale, fostering faster AI-driven innovation across the enterprise with security, scalability and compliance. In this blog post, I will walk you through end-to-end low-code machine learning with Ludwig and SAP AI Core.
... View more
a week ago
Posted in Artificial Intelligence and Machine Learning Blogs
Are you ready to dive into the latest SAP Business AI innovations shaping the future of enterprise technology? SAP TechEd 2024 is right around the corner, bringing together developers, IT leaders, and the broader SAP community to share insights, learn, and build solutions that will drive tomorrow’s businesses. Whether you're interested in artificial intelligence, cloud solutions, or next-generation tools, SAP TechEd has something for everyone. What to Expect SAP TechEd offers a comprehensive selection of sessions tailored to meet the needs of developers, technology enthusiasts, and business leaders. These sessions provide valuable insights into SAP’s cutting-edge technologies and their impact on industries across the globe. It’s a unique opportunity to engage with experts, developers, and IT leaders to learn new ways to advance your career and contribute to your company’s success. Discover how innovations such as Joule, SAP’s generative AI copilot, and large language models are transforming all business areas. Hear how SAP Business AI can enhance decision-making and agility while delivering tangible results. Key SAP Business AI Sessions to Watch KEY100CEST | SAP TechEd in 2024 Launch - Where ideas get real At SAP TechEd, experts, developers, IT leaders, and our SAP Community gather to learn, share, and build together. Hear firsthand from SAP Executives Muhammad Alam and Michael Ameling how innovations in SAP Build solutions, SAP Business AI, SAP BTP, and other solutions across our end-to-end portfolio are revolutionizing the tech world. AI100 | Artificial intelligence that creates real-world results Dive into the transformative potential of SAP Business AI and how you can use it in your daily work. Get an overview on the latest capabilities, and gain insights into more securely integrating AI into business applications, efficiently developing generative AI scenarios, and deploying the Joule copilot in your SAP software landscape. Explore essential capabilities, and elevate your AI strategy and business processes. Discover how SAP's advanced AI technologies can enhance your innovation. PAR103 | Unlocking AI innovation with SAP and Microsoft solutions Get ready to realize the full potential of AI innovation with SAP and Microsoft solutions. Explore the synergy between SAP Business Technology Platform (SAP BTP) and Microsoft Azure AI, explore groundbreaking ways to integrate SAP S/4HANA with Microsoft Copilot, and discover how you can harness the power of Azure AI with SAP solutions. Elevate your technical expertise, and learn how to drive actionable insights and business innovation with game-changing AI technologies. AI103 | Unlock generative AI more securely across your business with leading LLMs Learn how to more securely and smoothly leverage market-leading large language models (LLMs) to enhance productivity, creativity, and efficiency throughout your organization. Discover how to integrate and orchestrate these models programmatically using the generative AI hub and other advanced capabilities in the SAP AI Core infrastructure and SAP AI Launchpad on SAP Business Technology Platform (SAP BTP). AD180 | Build extensions with ABAP Cloud and Joule's ABAP developer capabilities Learn how to develop transactional SAP Fiori apps in ABAP Cloud using ABAP RESTful application programming model supported by generative AI. Explore features such as business events for loosely coupled integration scenarios and read-only "treeviews" for displaying hierarchical data. ABAP Cloud is our comprehensive development model for building clean-core-compliant apps, services, and extensions on SAP Business Technology Platform and SAP S/4HANA, both in the cloud and on premise. AI201 | Development with our generative AI hub, SDKs, and orchestration features Learn to develop dynamic generative AI applications using large language models (LLMs) with the generative AI hub in the SAP AI Core infrastructure. Understand how to take advantage of powerful SDKs and orchestration capabilities to accelerate your time to market. AD200 | Extend Joule with SAP Build and develop custom enterprise skills Discover the powerful synergy between Joule technology and SAP Build solutions that facilitates a smooth extension-building process. Learn how to enhance AI-powered capabilities in Joule with custom-developed enterprise skills tailored to your company's specific needs. By integrating with both SAP and third-party applications, Joule can instantly apply its newfound knowledge within SAP systems. Make your company more adaptive, responsive, and successful in the digital age. AD107 | Boost your coding efficiency: Explore Joule's ABAP developer capabilities Discover Joule's ABAP developer capabilities. Hear about detailed use cases, the generative AI road map, and insights on how these advancements can boost your productivity as a developer in your daily tasks. AI105 | Examining how to onboard and extend Joule Drive efficiency and productivity across your business by onboarding and extending Joule, our generative AI copilot, across your cloud solutions from SAP. Learn how to integrate Joule into SAP S/4HANA and SAP SuccessFactors solutions. Gain insights into the Joule road map, architecture, and preconfigured capabilities. Additionally, explore how Joule uses document grounding for more precise business answers, and learn how to extend this feature with your own HR policy document. DT109 | Explore AI-driven ERP transformation with SAP Signavio solutions Discover the future of SAP S/4HANA software transformations with the combined power of business process management and generative AI, leveraging the vast process data unlocked by SAP Signavio solutions. Our advanced process models enable quicker, simpler modeling, mining, and transformation, paving the way for innovative business optimization. Join us in pioneering the next era of SAP S/4HANA Cloud solution transformations. AI107 | From chaos to clarity: Boost your business data with SAP AI Services Discover how to maximize the potential of structured business data with SAP AI Services, such as Data Attribute Recommendation and Personalized Recommendation. And see firsthand how the Document Information Extraction service can revolutionize the handling of unstructured data right from the core input channels. The services act as a universal gateway, efficiently selecting and initiating relevant processes and integrating extracted results with structured data. AI106 | Harness retrieval-augmented generation in Joule and generative AI hub Gain a comprehensive understanding of retrieval-augmented generation (RAG), a cutting-edge approach to natural language processing that significantly enhances the accuracy of generated text. Explore our innovative RAG architecture within the Joule copilot and the generative AI hub capability, and learn how to implement RAG to provide relevant, context-rich answers from your large language models (LLMs), effectively minimizing AI hallucinations. AI101 | Enable more secure and responsible generative AI Gain an in-depth exploration of SAP's ethical and responsible AI practices. Drawing on an overview of our AI ethics guidelines and procedures, we explore how developers can integrate AI ethics into their use cases. Through practical examples, find out how features such as content filtering and data masking within our generative AI hub can support the responsible use of AI, helping ensure adherence to ethical standards and practices. AI180 | Build your own generative AI scenario for your use case Get hands-on with the generative AI hub in the SAP AI Core infrastructure, and leverage market-leading large language models (LLMs) to build your own use case. You can take advantage of templating, content filtering, and powerful SDKs to accelerate your project. Walk through the technology components and methodologies, with our support each step of the way. Learn how to efficiently create robust generative AI scenarios. AI194 | SAP Business AI and Joule: What’s in it for innovators? Discover how SAP Business AI empowers partners and customers to innovate and create value. We'll showcase the latest advancements in the generative AI hub and Joule capabilities and feature Agilita's experience using generative AI hub to develop high-impact solutions. AI102 | Enhance business predictions with SAP Foundation Model and knowledge graphs Discover how SAP Foundation Model, a table-native AI model, can accelerate time to value for predictive tasks on tabular data, addressing the limitations of large language models in business forecasting. Explore the integration of SAP Foundation Model and knowledge graphs, and see how explicit business semantics enhance business predictions. Register Now! Don’t miss the chance to attend these and many more sessions live. Register for SAP TechEd 2024 today and be part of the conversation that shapes the future of Business AI. Click here to register. Watch On-Demand Missed a session? Don’t worry—following the live event, all sessions will be available on-demand so you can catch up on all the action at your convenience. Keep an eye out for the on-demand links post-event.
... View more
2 weeks ago
Posted in Artificial Intelligence and Machine Learning Discussions
Hi Gurus, I am looking for a blog, reference document which can help me find a potential use case using AI ML offerings over BTP platform. I have already tried PO and PO confirmation POCs using DOX service. Could you please help me find an another typical use case in a business where we can leverage these AI ML service offerings of BTP and and make a real Impact. #AIML #SBPA #BTP
... View more
3 weeks ago
Posted in Artificial Intelligence and Machine Learning Discussions
Participant
1,252
3
1
Given my background in Ariba and AI, I decided to combine both skills to make a very simple PoC combining SAP Ariba, SAP Business Network and SAP BTP's SAP AI Service. In this article, you will find a use case where a buyer has generated a purchase order in SAP Ariba and sent it to the supplier via SAP Business Network. After receiving it in SAP Business Network, the supplier needs to extract the PDF information from this purchase order in order to generate a JSON message to integrate into their sales system (e.g. Salesforce). Here are the steps: Step-by-step Giving Ariba context 1. Through SAP Ariba Buying platform, buyer creates a PR that is converted in Purchase Order after all approvals: 2. The supplier has chosen SAP Business Network as the method of receiving purchase orders. The supplier receives this PO via SAP Business Network including PDF smartform: Using Document Information Extraction (SAP BTP AI Service) The following steps will then continue with the recognition of the fields of this PDF through the SAP AI Service of Document Information Extraction. 3. Through Document Information Extraction (SAP BTP), we generate the schema of the document to be recognized. In this case, we will use a pre-built schema provided by SAP. 4. Once the schema has been defined, the PDF is uploaded to the application. 5. After pressing ‘Confirm’ tile, the system starts with the recognition of the fields and the generation of the information through GenAI. 5. Finally, the system processes PDF and we extract the results: Human-in-the-loop is very important to verify that everything is correct. To facilitate this action, the system returns the percentage of possible error through the use of traffic lights (red, orange or green): Additionally, DIE is able to generate a document in different formats (text, json or csv) that we could use to integrate this order in a SAP or non-SAP application (e.g. Salesforce): Conclusions After this short PoC, we can see the potential of artificial intelligence in SAP through SAP BTP. Here are my thoughts: The pre-build schema represents a very important back-bone to start a more robust module for our clients. The use of the artificial intelligence tool is simple and intuitive. The application has been very easy to configure thanks to the SAP BTP boosters. The system allows us to customise any type of document we receive, so its flexibility and adaptability is very high. The extraction of the information in different formats allows the automation of future steps, such as, for example, the creation of an output of a warehouse, sales order, invoice, etc. We hope you find it interesting how we can combine the different solutions of the SAP offering so that through AI we can automate our processes and humans can perform tasks that bring more value to the business.
... View more
2024 May 2910:31 AM
Posted in Artificial Intelligence and Machine Learning Discussions
Introduction In today’s fast-paced world, understanding our eating habits and sensations can provide valuable insights into our overall well-being and improve our performance in work. Inspired by concept from “The Personal MBA by Josh Kaufman” about how tracking your food sensations could improve a little bit your performance in work and eager to test AI capabilities from SAP Joule, this blog post will guide you through the process of creating a CAP (Cloud Application Programming) application designed to record and track eating sensations. The main goal of this proof of concept is to demonstrate that functional users, even those without extensive coding backgrounds, can effectively generate and comprehend code using powerful AI tools like SAP Joule and Microsoft Copilot Webbrowser. By leveraging these advanced technologies, we aim to bridge the gap between functional expertise and technical implementation, empowering more individuals to contribute to innovative solutions. Join me as we explore the steps to build this application, from initial setup to deployment, and discover how AI can enhance our ability to create meaningful and user-friendly applications. Creating and translating requeriment using Microsoft Copilot Creating a requirement in your native language and then translating it can be much easier than doing it directly in English. In this case, I decided to add to the prompt to simplify and make the requirement much more understandable for later programming. The quality of the documentation and its understanding is essential for artificial intelligence to program correctly. Once we have the requirement, we will set up an SAP BTP account to test creating a CAP application using SAP Build and SAP Joule without extensive knowledge of these technologies. IMPORTANT: In this test, we will only program using natural language. Set-up SAP Build Code using boosters First of all, we need to get the SAP Build application running on our SAP BTP account. In this case, I don’t have technical knowledge in SAP BTP, but we can use boosters to have a fully operational SAP Build programming environment without needing extensive knowledge. Setup SAP Build Code in SAP BTP | SAP Tutorials Special thanks to Paola Laufer for the tutorial. Programming application using SAP Joule and understanding details with Microsoft Copilot IMPORTANT: Highlight that the steps used were taken from material provided by SAP and adapted to my application. Once again, special thanks to Paola Laufer for this SAP Build Code Test Drive that has served as my guide. Create a Full-Stack Application with SAP Fiori UI using Joule in SAP Build Code | SAP Tutorials Let’s get started with our application! Artificial intelligence always needs to be reviewed and adapted by humans. Although this could probably have been done through a series of prompts, I have adapted the requirement to a format that I believed SAP Build would understand better. Create Data Entities with SAP Joule Using the requirement and the SAP Joule option in SAP Build Code, we can specify that we want to create a CAP application and generate the necessary data models and services. 1. Prompt SAP Joule with requirement: 2. Accept the response: 3. Results: Enhance the Application Data Using SAP Joule The data for testing applications must be real and reflect reality. Generating these datasets can be a tedious task that consumes a lot of time and imagination. You’ll never have to do this again because SAP Joule can generate the data you need to work on your application! 1. Prompt to generate data: 2. Accept and trigger the information for your testing table: 3. AI makes mistakes, but we review and correct them by filling in some fields that were left empty: Sorry my mistake here, I should request to SAP Joule to delete 'minutes' column because it's duplicated with 'hour'. Create Application Logic with SAP Joule Now it’s time to add logic to our application to ensure that the behavior of the fields is as expected. 1. Add logic: 2. Prompt SAP Joule with data validation logic that you want to perform: 3. SAP Joule triggers a proposal of the code: 4. As I'm a noob in CAP language, I've requested to Microsoft Copilot to ensure that all steps are good and to understand the code: 5. Once code is reviewed and understood that works fine, this code is accepted and trigger to the application: Add UI to the Application We have all the code now; it’s time to create the screens to enter and view the information. Let’s do this! 1. Create UI application: 2. Most important step here is to decide which actions will be performed by this application. In this case, we've selected 'Form Entry Object Page' because UI application will be responsible to collect data from user: 3. UI application is created with few easy steps: 4. Repeat steps for application which is responsible to show information. In this case, we need to select 'List Report Page': 5. Finally, we can show how both applications are created: Run application Everything is ready! Now let’s run the application and test it. Overview about 'Food tracking for your job performance' application Application screen The requirement called for two applications: one to add food items and another to review the data. We have successfully met this requirement: Add a new entry The requirement called for help fields to enter the data and only allow certain values. The day and time fields have the correct drop-down menus. But it wasn’t able to create a drop-down menu for the ‘Type of food’ and ‘Transmitted Sensation’ fields (it probably could have done this by iterating the system). Despite not having a selection menu for ‘Type of food’ and ‘Transmitted Sensation’, the system was able to create validations to ensure the correct field is selected. Show your data The system was able to generate the table to display and it works correctly, showing the entry added in the previous step: Additionally, the requirement asked to group by ‘Day’, ‘Type of food’, and ‘Transmitted Sensation’, and this requirement has been met: Conclusion The application requirement wasn’t complicated, but I believe it’s enough to demonstrate that for simple application cases, we can use SAP Joule to generate a fully operational application. As we can see from the analysis, the application meets 90% of the requirements and only failed in generating the drop-down menu. For me, the conclusion is clear: without any CAP knowledge, I was able to create an application in less than 45 minutes and generate a blog in less than 2 hours. Now, I’m excited to see what I could achieve with a bit more time! End credits Time is very valuable and creating blog easy to understand and with high quality content are difficult with few time. AI can help us doing it and you have an example with this post which mainly created using Microsoft Copilot (AI). Time consumed for this PoC: 2.5 hours of a boring Sunday
... View more
3 weeks ago
Posted in Artificial Intelligence and Machine Learning Discussions
Participant
117
0
0
Introduction In today’s fast-paced world, understanding our eating habits and sensations can provide valuable insights into our overall well-being and improve our performance in work. Inspired by concept from “The Personal MBA by Josh Kaufman” about how tracking your food sensations could improve a little bit your performance in work and eager to test AI capabilities from SAP Joule, this blog post will guide you through the process of creating a CAP (Cloud Application Programming) application designed to record and track eating sensations. The main goal of this proof of concept is to demonstrate that functional users, even those without extensive coding backgrounds, can effectively generate and comprehend code using powerful AI tools like SAP Joule and Microsoft Copilot Webbrowser. By leveraging these advanced technologies, we aim to bridge the gap between functional expertise and technical implementation, empowering more individuals to contribute to innovative solutions. Join me as we explore the steps to build this application, from initial setup to deployment, and discover how AI can enhance our ability to create meaningful and user-friendly applications. Creating and translating requeriment using Microsoft Copilot Creating a requirement in your native language and then translating it can be much easier than doing it directly in English. In this case, I decided to add to the prompt to simplify and make the requirement much more understandable for later programming. The quality of the documentation and its understanding is essential for artificial intelligence to program correctly. Once we have the requirement, we will set up an SAP BTP account to test creating a CAP application using SAP Build and SAP Joule without extensive knowledge of these technologies. IMPORTANT: In this test, we will only program using natural language. Set-up SAP Build Code using boosters First of all, we need to get the SAP Build application running on our SAP BTP account. In this case, I don’t have technical knowledge in SAP BTP, but we can use boosters to have a fully operational SAP Build programming environment without needing extensive knowledge. Setup SAP Build Code in SAP BTP | SAP Tutorials Special thanks to Paola Laufer for the tutorial. Programming application using SAP Joule and understanding details with Microsoft Copilot IMPORTANT: Highlight that the steps used were taken from material provided by SAP and adapted to my application. Once again, special thanks to Paola Laufer for this SAP Build Code Test Drive that has served as my guide. Create a Full-Stack Application with SAP Fiori UI using Joule in SAP Build Code | SAP Tutorials Let’s get started with our application! Artificial intelligence always needs to be reviewed and adapted by humans. Although this could probably have been done through a series of prompts, I have adapted the requirement to a format that I believed SAP Build would understand better. Create Data Entities with SAP Joule Using the requirement and the SAP Joule option in SAP Build Code, we can specify that we want to create a CAP application and generate the necessary data models and services. 1. Prompt SAP Joule with requirement: 2. Accept the response: 3. Results: Enhance the Application Data Using SAP Joule The data for testing applications must be real and reflect reality. Generating these datasets can be a tedious task that consumes a lot of time and imagination. You’ll never have to do this again because SAP Joule can generate the data you need to work on your application! 1. Prompt to generate data: 2. Accept and trigger the information for your testing table: 3. AI makes mistakes, but we review and correct them by filling in some fields that were left empty: Sorry my mistake here, I should request to SAP Joule to delete 'minutes' column because it's duplicated with 'hour'. Create Application Logic with SAP Joule Now it’s time to add logic to our application to ensure that the behavior of the fields is as expected. 1. Add logic: 2. Prompt SAP Joule with data validation logic that you want to perform: 3. SAP Joule triggers a proposal of the code: 4. As I'm a noob in CAP language, I've requested to Microsoft Copilot to ensure that all steps are good and to understand the code: 5. Once code is reviewed and understood that works fine, this code is accepted and trigger to the application: Add UI to the Application We have all the code now; it’s time to create the screens to enter and view the information. Let’s do this! 1. Create UI application: 2. Most important step here is to decide which actions will be performed by this application. In this case, we've selected 'Form Entry Object Page' because UI application will be responsible to collect data from user: 3. UI application is created with few easy steps: 4. Repeat steps for application which is responsible to show information. In this case, we need to select 'List Report Page': 5. Finally, we can show how both applications are created: Run application Everything is ready! Now let’s run the application and test it. Overview about 'Food tracking for your job performance' application Application screen The requirement called for two applications: one to add food items and another to review the data. We have successfully met this requirement: Add a new entry The requirement called for help fields to enter the data and only allow certain values. The day and time fields have the correct drop-down menus. But it wasn’t able to create a drop-down menu for the ‘Type of food’ and ‘Transmitted Sensation’ fields (it probably could have done this by iterating the system). Despite not having a selection menu for ‘Type of food’ and ‘Transmitted Sensation’, the system was able to create validations to ensure the correct field is selected. Show your data The system was able to generate the table to display and it works correctly, showing the entry added in the previous step: Additionally, the requirement asked to group by ‘Day’, ‘Type of food’, and ‘Transmitted Sensation’, and this requirement has been met: Conclusion The application requirement wasn’t complicated, but I believe it’s enough to demonstrate that for simple application cases, we can use SAP Joule to generate a fully operational application. As we can see from the analysis, the application meets 90% of the requirements and only failed in generating the drop-down menu. For me, the conclusion is clear: without any CAP knowledge, I was able to create an application in less than 45 minutes and generate a blog in less than 2 hours. Now, I’m excited to see what I could achieve with a bit more time! End credits Time is very valuable and creating blog easy to understand and with high quality content are difficult with few time. AI can help us doing it and you have an example with this post which mainly created using Microsoft Copilot (AI). Time consumed for this PoC: 2.5 hours of a boring Sunday
... View more
3 weeks ago
Posted in Artificial Intelligence and Machine Learning DiscussionsTop Kudoed Posts
| Subject | Kudos |
|---|---|
| 2 | |
| 2 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 |
Top kudoed authors
| User | Count |
|---|---|
| 4 | |
| 2 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 | |
| 1 |
Members (1,216)
| User |
|---|
manish_gangwal
Discoverer
|
DilipKumarR
Newcomer
|
Bas123
Explorer
|
Nico
Explorer
|
DarCalFer
Newcomer
|