Pravin Gadade

A successful handover plan for an AI solution becomes critical to ensure sustainable solution / automation, its adaptability and continuous value generation in future. Having defined handover process at organizational level can be a good practice to follow. This will include: Checklist: Having standard handover checklist and approval matrix will be helpful to avoid any Standard Operating Procedure (SOP): A SOP should be prepared at detailed level for the end users and the administrator. This should be treated as go to document when any users find any questions or have any queries at any given point of time. Make sure to have standard FAQs are included to make it more effective. Solution Architecture: Prepare a blue print of solution architecture at detailed level including integrations points. Maintenance Schedule: Guide business owner / administrator about periodic updates in KB and external links. Trainings: Schedule periodic pre-defined trainings. Monitoring: Alerts for anomalies and observations. Issue logging system and setting severity levels and escalations.

I would like to consider an example of an AI Agent handling employees queries to answer this requestion. Payroll queries for high-stake and very sensitive. If not addressed properly, it may directly impact on employees morale and unwanted escalations. This agent includes below steps: Understand nature of the query (Under payment, excess tax, leaves, bonus, etc.) Verifying KB (ERP, HRM Tool, Bank statement) Communicate responses Ensure compliance Techniques to keep AI interaction focus and efficient: Vague framing: “Let me help you on your question by referring database.” Role- based framing: “As your payroll assistant, I will guide you on underpayment issue and escalating if needed.” Structured Flow: Identifying issue and reconfirm from the user. Collect data, verify compliance from KB and summarize Escalation path as per the nature of query Resolution and follow up on the query Insert confirmation check point, e.g. “Is this accurate before we proceed?” Context retention by referring to the conversation in the past. Prompt should include emotions and avoid unwanted verbiage unless asked. Keep taking feedback and rating on user experience.

Problem Statement: Extracting more than 400 invoices from group mailboxes is manual task and time consuming. I am planning to introduce this use case in my organization using AI solution. This more popular AI solution introduced by many organizations. With the help of no code or low code AI available solutions, invoices can be extracted from group mailboxes and saved on OneDrive. It can also help AP team to extract invoice details like vendor name, invoice number, date, amount and save this data on excel spreadsheet for review on daily basis. Quality of the prompt will have direct impact on the tool operation and quality of the data. If they are not properly used then it will not will not derive expected business results and dissatisfaction in end users and management. Vague prompt: “Extract invoices form the shared Mailbox [[email protected]] and save them in in OneDrive [Path] on daily basis at 8:00 AM. Collect invoice details and update spreadsheet saved on One Drive [Path] ” Improved Prompt: At 8:00 AM daily, scan the shared mailbox [[email protected]] for emails received after 8:00 AM on the previous day. Filter Criteria: The email subject line must contain the keywords: “invoice” or “bill”, OR The email must include a PDF attachment that appears to be an invoice. Actions: If the email does not meet the above criteria, mark it as unread. If the email qualifies, move it to the [AP Invoices] folder within the group mailbox. Extract the PDF attachment and save it to OneDrive at the following path: [Path]. From each PDF, extract the following details: Vendor Name Invoice Number Invoice Date Invoice Amount Save the extracted data into an Excel spreadsheet located at: [Path]. Notification: Once the task is completed, send a summary email to [[email protected]] with a list of vendor invoices saved to OneDrive.

Periodic updates and maintenance strategy becomes the integral part of implementation and control strategy while implementing any AI solution. Without monitoring strategy accuracy and compliance can suffer. Lets consider below failure scenario from Accounts Payable Process: An AI solution is used to extract and validate invoice data like vendor name, amount, date, etc. from email and pdf copies and save those copies on OneDrive designated folder. Over the period of time performance may degrade due to various factor like: Prompt updates not considered for new invoice templates, formats changes in business requirements. Input scenarios not updated periodically. Changes I workflow like changes in approval flow and vendor policies Early Warning Signs: Increase in manual correction by the AP staff Payment delays Vendor complaints Mismatch between PO and invoice data Drop rate / error rate in processing Monitoring: Set alert if drop or error rate drops by 10% and review scenarios Trak manual corrections using tracker Monitor trends Monitor complaints from vendor and AP staff

Being a AI Solution Architect, it becomes crucial that design solution which will remain sustainable and clear roles and responsibilities are assigned. This approach can be defied with the help of below points: Indicators: Exceptions: Review audit logs to understand occurrence of exceptions and the output generated by AI solution. If exception numbers are high with unexpected outputs then it clearly suggest that it is not aligned with the business requirement. User Experience: Bad user experience and not using the AI tool will indicate it is not aligned with the requirements. Misalignment with policies, SOPs, business updates – If AI solution does not take care of frequent or periodic updates then it will not be able to function effectively. Preventive Actions: Governance: Define roles and responsibilities to ensure solution remain updated all the time. Define timelines. Feedback: Ensure timely feedback from the end users to understand ongoing issues. Reporting: Ensure periodic updates shared with the respective teams. Change Management: Establish change management framework and align particular AI solution with it. SOPs: Updating SOPs on periodic basis or as and when major changes are implemented.

Being a transformational professional it becomes crucial to understand VOC, CTQ, Target business impact (time reduction, cost saving, generation of revenue). Below scorecard framework can be used to ensure setting priorities to the AI project solutions. I have never worked on such kind of initiatives hence do not have examples to share from the past experience. AI Scorecard Framework. They will be rated out of 10. Highest score project will be prioritized. Problem and business impact ROI – What is possible time & cost saving, revenue generation in numbers. Rate criticality of the problem and current challenges. Strategic Alignment: Goals of the organization – Digitization, compliance, improve customer experience. Rate as per the priority. Rating on readiness of accepting failure. This will determine project impact on business if it does not work the way it was planned. Infrastructure: Pre-implementation: Is data available and as per the requirement of the proposed AI solution. Cost Effectiveness: Cost for implementing AI solution meets ROI requirements in point # 1.a.

My organization is facing challenges in managing Accounts Payable process. There are instances where we ended up missing out on renewal of contracts, invoice booking, late payments to vendors, data entry errors, manual email tracking, etc. They have direct impact on our business and poor experience to maintain vendor relationship. We are looking or cost effective AI tool that can help us to reduce these errors and bring efficiency. Application of Swiss Cheese Model: Slices of Cheese (Defense Layers): Contract Controls: Make sure contract are active with defined term, scope and payment terms. Invoice Receipt, 3 way match and data entry in ERP: Ensure all invoices received are considered, matched and booked. Update invoice details in log. Approval Flow: Ensure all invoices are approved by the respective department head and establish standard follow up / escalation mechanism. Payment Processing: Paying in batches with approvals from the respective heads from the treasury. Vendor Reconciliation: Perform vendor reconciliations by assigning risk levels to the vendor in high, low and medium by assigning frequency of the reconciliation. Holes in the Cheese (Potential Weaknesses): Manual Group Email Management: Oversight to read email, deletion, erroneously marking an email as completed. Data Entry: Incorrect vendor name, amount, payment date, service descriptions, account number, double invoice processing. Approvals: Lack of timely approvals from the department heads. Unclear responsibility: Assumptions / miscommunication between procurement team and invoice processing team and approvers. Visibility: Lack of periodic dashboards for management review. Process Excellence: We are planning to introduce AI solutions to minimize manual efforts, bring efficiency and improve vendor experience. Solution: Email Parsing and Classification with Power Automate + AI Builder. Phase 1: Develop solution and BRD approval from the management Phase 2: Perform trial with limited emails and corrections Phase 3: Train operations team Phase 4: Implement Phase 5: Governance

Change management is critical factor for the success and failure of any project irrespective of its scale level. A project may fail if it has not been properly discussed, communicated and adopted by the involved stakeholders and the project team. Relevance of change management in DMAIC roadmap: Define: Clearly defining project scope and problem statement with the stakeholders. Engaging stakeholders in early stage of the project becomes essential to build trust and alignment. Example – Reduce invoice processing time – In this scenario involving finance manager to understand pain points becomes essential. Measure: Gathering data from the respective stakeholders and getting it validated. Example – Coordinating with the AP team to ensure base line data and relevant matrices. Analyse: Identify root cause of the problem statement. Example – Initiate brainstorming sessions with the cross functional teams like finance, IT and vendors. Improve: Design solution. Example – Build prototype and pilot. Example – Train AP super users on changes and run pilot. Control: Sustain changes. Example – Train all users / teams. Incorporate changes in SOPs and monitor matrices.

The primary focus of Business Excellence Driven Organization is to keep innovating and improving processes in order to boost efficiency and add value to the organization. However use of AI solutions while delivering results has resulted into the emerging need of governance. This can be evaluated with the help of below points: Fairness / Biasness – AI solutions should be implemented without any bias in data, people, department, etc. There should be enough testing performed and reviewed on an AI solution output to verify the authenticity of the outputs. Transparency – AI data flow should be understandable for the stakeholders. There should be documents in place narrating how AI modules are trained, data inputs and desired data outputs. Allocation of duties & roles – Assign roles to the individuals with defined accountabilities & responsibilities. This will also include assigning risk management roles and periodic ethical reviews from the auditors. Examples, Chief AI Officer, AI Solution Engineer, AI Risk Management Officer, etc. Privacy – Implement policy for data protection getting used for AI solutions’ lifecycle. Implementation of robust cyber security measure becomes integral part of the process. Sustainable Solutions – AI solutions should remain evolving with the time to avoid any negative risk on outputs. Formation of cross functional governance committee for better implementation and keep involved all the departments within organization. Periodic evaluation of talent, readiness of the solution, data quality, budget, etc.

Process can be improved when: 1. There are repetative tasks and standard process exists withiout significant exceptions or deviations in the tasks. 2. There are lot of manual intervention required which can be processed or automated with the help of a tool. 3. Process tasks are measureable in terms of time, savings, etc. 4. They are rule based which can be understood. Example: Automating bank reconciliation process by applying formulas / VBA Macros to eliminate maunal tasks of matching trasactions one by one. Process can be re-imagined when: 1. It is complex in nature and lot of exceptions are involved. 2. When there are dependant real time sources involved while performing tasks. 3. When existing process does ot meet the expectations / required / correct output. 4. When performing tasks by a human is significantly time consuming with copromise on quality. 5. When external enviorment / market is evolving but processes we are following are out dated. Example: Market analysis report before launching a new product which can be comparitively easier to prepare, less time consuming, more detailed as comapred to prepared by a consultant.