Menu
Trusted technology partner to some of the world’s most successful private equity funds and their portfolio companies, through the investment cycle.
Commencing an AI roadmap
Regardless of where a company is in the usage of AI it is timely to understand that many core business concerns and functions are now capable of significant augmentation through AI approaches. The rate of AI adoption is unprecedented and offers compelling ROI for businesses positioned to adopt these. The D-LAB network has access to a rapidly maturing set of use cases for core business functions where real-world experience and expertise can be applied with minimal risk and where ROI can be delivered in a 3–6-month timeframe.
Key business considerations
- AI Use cases need a robust view of the following factors prior to commencing any pilot or POC AI roadmap project :
- Time to Market: The use cases indicated below are all feasible for middle-market organisations. Typically, these should be achievable in 3-6 months (to an MVP) level.
- Return on Investment: ROI is ultimately dependent on business circumstances. AI-delivered benefits are typically 20-50+% range when factored against the un-augmented business function
- Risk of Development: The D-Lab network can identify and deliver real-world solutions for use cases where a repeatable capability already exists to significantly reduce the risk of development.
- Business impact : Beyond ROI and finesse business benefits typically include enhanced reach and speed of business activities.
Five things to understand about AI.
- AI is about data. Effective AI requires a sufficiency of high-quality data. Data drives AI. No Data - no insights, no possibilities. Data cannot be manufactured; it must come from the real world.
- AI delivers interpretations and probabilities, not 100% certainties. AI-driven processes must be engineered to manage this.
- AI encompasses many evolving methods. AI in several forms has been evolving for the past couple of decades. There are a multitude of AI techniques and approaches being used today, evolving at different rates, and with different strengths and specialisations.
- AI Solutions are driven by specific objectives. The sophistication and complexity of AI Solutions is reflected by the width of their objectives. Defining requirements is as important as ever.
- AI solutions can improve over time. Well-engineered AI solutions grow and learn as they are used, becoming more accurate at their specific tasks.
Use cases for common business functions
Within the business functions considered in this document, there is considerable overlap E.g., Operations and Analytics span and influence many of the others.
AI Techniques such as anomaly detection, the use of chatbots, etc, also overlap many of these business functions but are applied to the specific concerns of individual functions.
These commonalities present an opportunity to replicate expertise over multiple business concerns in adopting particular AI-based approaches.
- IMPACTS
- Key AI benefits are called out for each business function,
- Agility – Function cadence and responsiveness
- Scale – Function operational capacity and throughput
- Efficiency – Function cost and resource impact
- Depth – Function finesse, subtlety and reach.
- Other compelling benefits are indicated as shown.
Core business functions with their major AI impacts
Current state AI for business intelligence
- Current approaches such as Data Lakes provide analytics with machine learning capability along with element or keyword-based focus. Relationships between data nodes are discoverable by machine learning methods.
Potential BI data sources
- Enhanced AI capabilities can now mine and merge a wide range of data types and sources to provide augmented business intelligence. This increased capability is additionally more agile permitting a more responsive level of real-time analysis and insight.
- Conversational Data – data from business engagements and customer interactions can be analysed to extract valuable insights.
- E-Commerce – timely and deep insights can be extracted from raw e-commerce data using ML-based insights merged with other data sources and triggers.
- Geo Analytics – Geo operational data can be used to identify trends and identify potential high-cost tuning/optimisation for concerns such as transport fleets and logistics.
- Image Recognition – Image data can now be mined and interrogated to provide observational insights to augment KPI reporting in manufacturing, construction, and other site-based activities.
AI trends for BI and their business impact
- Increased finesse – Generative AI, through its language capability, allows BI to consider additional factors influencing the data being examined, magnifying the depth and subtlety of BI
- AI-driven ROI measurement Businesses are increasingly looking for ways to measure and improve their return on investment (ROI). Using AI initiatives assists in the definition of clear goals, metrics and benchmarks for business activities projects, as well as tracking and evaluating their outcomes and impacts. AI-augmented ROI management allows businesses identify and prioritise the most valuable business initiatives, justify spending and demonstrate value to stakeholders.
- AI-augmented automation: BI is often an operational challenge for organisations where data must be collated and prepared from many sources, often against tight deadlines. Using contemporary AI methods allows for the automation of these activities permitting more complete, efficient, and responsive BI pipelines.
Use cases: AI can enhance current analytics and BI/MI by improved “joining of the dots”, leveraging new and meaningful data held in disparate silo’s and in automating the collation and alignment of data for analysis. AI can help automate and enhance the measurement and reporting of KPIs, making them more accurate, timely, and actionable.
AI-augmented performance data measurement – Using AI to gather richer and widely related data for KPI measurement.
KPI anomaly detection – Modelling data and KPI’s to detect anomalies. Detecting data quality issues.
Predictive analytics – Using predictive methods to anticipate KPI’s against expected trends. Identifying false positives and false negatives.
Improved strategic alignment – AI driven interconnected and dynamic KPI’s to monitor and improve strategic alignment.
Governance KPI’s – the increase in fidelity and coverage for data underpinning business activities allows the derivation of Governance KPI’s
Enhanced data consolidation – Mapping and interpreting data from disparate sources to form consolidated data views for KPI gathering.
KPI Frequency improvements – Using AI driven data collection to improve KPI frequency and granularity.
BI verification – crosschecking and triangulating results by using automated code generation against BI requirements.
Better self-service KPI’s – Allowing unstructured KPI’s to be composed and interrogated through self-service interfaces for business partners and customers.
Current state AI for marketing
- AI is currently supporting marketing activities in several ways :
- Search Engine Optimisation (SEO): AI is being used to conduct keyword research and develop optimised content, improving search engine rankings and increasing website traffic.
- Chatbots: AI-powered chatbots handle customer support improving customer engagements. Chatbots handle simple queries, reducing impact on human agents.
- Personalisation: AI is able to personalise content and offers based on user data and behaviour patterns, increasing customer engagement and increasing conversion rates.
- Predictive Analytics: AI-powered predictive analytics forecasts future trends, identifying potential growth opportunities for growth, helping businesses to make data-driven competitive decisions.
Forward AI trends for marketing and their business impact
- Generating content: AI creates more engaging and relevant content for their target audience, by analysing data on customer preferences, behaviour, and feedback. AI generates draft headlines, captions, slogans, advertising copy, and entire articles using natural language generation (NLG) techniques.
- Monitor consumer behaviour – AI quickly picks up on changes in consumer behaviour: helping marketers adapt to changing market trends and expectations, Using predictive analytics and sentiment analysis AI measures and forecasts customer demand, satisfaction, and loyalty. AI also aids optimises campaigns and strategies based on real-time feedback and performance metrics.
- Predicting customer spending patterns – AI can help marketers increase revenue and profitability, by using ML models to predict customer lifetime value (CLV), churn rate, and cross-selling opportunities. AI can also help marketers set optimal prices and discounts based on customer behaviour and demand.
- Performing market research – AI can help marketers gain insights into their competitors, customers, marketplaces and industry trends, by using web scraping, text mining, and image recognition to collect and analyse data from multiple sources. AI driven insights can identify gaps and opportunities in the market..
Use cases: Marketing based AI use cases are increasingly driven from a more refined high-fidelity understanding of both target customers and the business marketplace. The ability of contemporary AI to determine highly actionable data based on the discovery of high impact relationships and attributes of the underlying data is key to realising major benefits from AI driven marketing activities.
Know customer preferences – AI can derive customer preferences from different sources (ERP, social media, web analytics, surveys, CRM, etc.). Focused strategies and campaigns can be created by clustering customers on key data (such as demographics, interests, behaviours, etc.)
Predictive customer and marketplace analytics –AI helps marketers make data-driven decisions, by using AI to analyse customer and market data and provide actionable insights.
Intelligent Recommendations – AI assists marketers increase conversions and retention by using recommender systems to suggest products, services, or content relevant to customers. AI can also use natural language processing (NLP) and computer vision to understand customer queries and provide accurate answers or solutions.
Data infrastructure sophistication – By gathering and analysing data. AI can help marketers collect/process build sophisticated models from various sources, such as websites, social media, and 3rd party data sources.
Content Creation – Generating marketing content based on chosen themes can be achieved with Generative AI to a very high level of sophistication in a fraction of the time formerly take by human content creators, tailored for customer traits.
Email Personalisation – Email personalisation can be achieved based on complex customer observations of their preferences, reviews and other interactions with a business to more form more precisely targeted marketing campaigns
SEO content optimisation – Ai based analytics of SEO trends can enable SEO tuning to magnify the visibility for web site content. AI generated content such as blog posts can also drive SEO precision yielding more efficient marketing and advertising spends.
Social Listening – AI based social media listening can uncover and highlight marketing trends as well as identifying target demographics and special interest groups for marketing purposes.
AI enhanced CRM – AI based data gathering and processing can be used to enrich CRM platforms with augmented details of customer browsing and buying trends.
Current state AI for operations
- Operations is already influenced by several differing usages of AI.
- Operational workflows – AI can be used to automate and streamline operational workflows, making them more efficient and effective.
- Supply Chain Management – AI can also help enhance supply chain communication by providing real-time responses, enriching customer communication and satisfaction by also offering order updates and shipment tracking information.
- Cybersecurity – AI is currently being used to strengthen Cybersecurity by monitoring for anomalies in usage patterns and by scanning for evidence of external threats.
- Customer contact – support and helpdesk operations are frequently supported by. Using chatbots to assist in front line support and manage high call volumes.
- Operational agility – Machine learning AI is frequently used to process large volumes of data quickly and derive key operational concerns more rapidly than in the past. The frequency of checks has been raised to days or hours from weeks or months resulting in a large boost to operational agility.
AI trends for operations and their business impact
- Process automation: AI can help automate repetitive, rule-based tasks that are prone to human error or inefficiency. For example, AI can be used to automate invoice processing, data entry, quality control, inventory management, and more.
- Production optimisation: AI can help optimize production processes by analysing data from sensors, machines, and other sources. AI can also help with predictive maintenance, anomaly detection, energy management, and waste reduction.
- Cybersecurity and fraud management: AI can help protect businesses from cyberattacks and fraud by detecting and preventing malicious activities, such as phishing, malware, identity theft, and more. AI can also help with risk assessment, compliance, and audit.
- Customer operations: AI can help improve customer-based operations by providing personalised and timely responses to customer queries and requests. AI can also help with chatbots, voice assistants, sentiment analysis, and feedback collection.
- Content production: AI can help create content such as text, images, videos, music, and more. AI can also help with content analysis, summarisation, translation, and editing.
Use cases: Operations based AI use cases are being driven from a higher frequency and diversity of operational data. This diversity feeds into many data-driven processes each of which can be examined or augmented by AI. All these use cases require that the relevant data can be made available.
Intelligent Automation – The combination of AI and Robotic process Automation allows AI derived actionable insights to be translated into automated responses and workflows, raising operational agility and increasing efficiency.
Facilities Management – AI can work in conjunction with building sensors and other automation to increase savings and efficiencies for facilities management
Predictive maintenance – Again working with sensor data and in conjunction with CMMS (Computerised Maintenance Management Systems) AI can predict equipment failure and proactively schedule interventions, preventing downtime, disruption, and lost revenue.
Incident root cause analysis – AI can assist in analysing data following an incident to determine root-cause, allowing pro-active remediations to be made.
Content generation – Operations itself routinely needs to publish process description, policies and other content. A suitably trained LLM application can assist in the generation of materials like these.
Fraud prevention – AI can analyse financial data to highlight discrepancies and anomalies, detecting and preventing fraud.
Document processing – document processing, such as invoices, policies, contracts, rental agreements etc can be scanned and processed by document-based AI systems to speed up processing and apply consistency.
Cybersecurity Anomaly and threat detection – AI can detect cybersecurity threats and anomalies, raising alerts and allowing proactive responses to be made
Inventory and Supply chain management Trends and utilisation patterns – Stock control and order management systems can use AI to predict usage patterns and anticipate supply chain characteristics to fine-tune order and inventory management.
Analysis of CX/UX patterns – analysis of usage patterns for applications can determine CX and UX patterns – this knowledge can be used to alter the applications to realise specific benefits
Virtual Assistants – Working with productivity tools such as MS 365 virtual assistants can manage priorities and streamline personal workflows.
Advanced cost analysis – AI can manage disparate costing information from service providers to optimise service spending.
Current state AI for sales
- AI has become a game-changer in the sales industry. According to a recent report by HubSpot, more than half of high-performing sales organisations leverage AI. Specific AI approaches are:
- Data Enrichment, Analysis, & Applications – AI tools can help sales teams access advanced data enrichment, analysis, and applications to improve their performance.
- Conversation Intelligence – AI-powered conversation intelligence software can enhance sales performance by analysing conversations and providing insights on how to improve them.
- Automated Lead Scoring – AI can help sales teams automate lead scoring, which is the process of evaluating leads based on their likelihood to convert into customers.
- Personalisation – AI tools can collect and analyse data points to provide insights that enable sales teams to personalise their approach to different personas.
- Adapting Quickly to Buyer Needs – AI can help sales teams adapt quickly to changing buyer journeys and increased buyer sophistication.
AI trends for sales and their business impact
- Froward trends in AI for sales focus on leveraging LLM data insights gathered from sales and CRM systems and other sources :
- Customized proposals – AI can collate customer pain points, industry trends, pre-sales conversations, etc., to create tailored proposals.
- Identification of upselling & cross-selling – Using historical and current data AI can predict which products or services are most likely to appeal to existing customers.
- Lead refinement – AI can create targeted advertising campaigns based on customer behaviour, preferences, etc., AI can also create customized video demos highlighting the value proposition of a product/service.
- Prioritizing best leads – AI can score leads based on their probable outcomes, potential value, urgency, etc. AI can also help sales focus on, and execute follow-up tasks.
- Identifying on-track/at-risk opportunities – AI can monitor the progress of sales activities and call out changes or risks that might impact outcomes, providing guidance and options on how to improve deal outcomes.
- Increasing forecast accuracy – AI can predict the revenue and pipeline of the sales team, identifying forecasting gaps and risks.
- Adapting sales strategy – AI can adjust sales strategy for specific customers and messaging based on the buyer’s history, persona, and intent.
- Content generation – AI can create personalized sales collateral for each prospect using generative AI.
Use cases: Sales based AI use cases are increasingly driven from analytics of customers sales outcomes and sales characteristics. AI permits high fidelity decision making at scale allowing large numbers of customers and sales diversity(SKU’s) as well as varied sales scenarios to be managed and conducted with consistency and precision.
Predictive pricing & pricing trend analysis – Utilising in-house and marketplace sales data to detect market trends and movements to drive pricing adjustments.
Sales forecasting – project sales forecast for actual and what-if scenarios against marketplace to co-ordinate planning activities and maximise opportunities
Lead Generation – Analyse marketplace data to identify or anticipate demand.
Sales response suggestions – process and categorise sales responses feedback to formulate sales team and customer specific actions.
Sales Strategy suggestions – Identify changes in customer behaviour and service/product to make sales strategy suggestions.
Sales lead preference analysis – evaluate sales lead priority metrics and perform sales scenario analysis to determine efficacy and maximise business advantage.
Social Listening – review social media for customers and customer organisations to determine customer preferences and demand.
Contact Analytics – monitor customer communications to determine customer preferences and demand and identify key decision makers.
Voice of Customer Analytics – monitor any VOC reviews and feedback on third party review sites or gathered through company websites and social media channels.
Competitor Review Analysis – Analyse reviews for competitor products and services to evaluate gaps and differentiators.
Sales Funnel Analysis – Analyse all steps of the sales funnel to identify problems ( e.g. Delay in follow-ups, Not enough sales capacity, Pursuing deals that are too small or poor fit, Low contact rates Inaccurate performance metrics, etc.)
Sales Support intelligence – identifying and communicating real-time sales trends to sales support.
Sales Performance analytics – AI streamlines the data collection, processing, and analysis of sales performance, leading to a more real-time to issues, trends and anomalies, leading to improved and rapid decision-making.
Current state AI for HR
- HR has often been a paper-based activity in the past. AI involvement so far has been light as there is a lack of the high-quality data needed to drive robust HR decision-making. Technology innovation for HR has centered around automating processes with self-service portals, mobile applications, and platforms for time sheeting and for HR administration and review processing systems such as Workday.
AI trends for HR and their business impact
- Recruiting: AI is helping companies find the right talent by analysing resumes and matching them with job descriptions.
- Performance reviews: AI can help managers conduct performance reviews by providing data-driven insights into employee performance.
- Employee onboarding/offboarding processes: AI can help automate the onboarding and offboarding processes by providing new employees with relevant information and training materials.
- Employee engagement initiatives: AI can help HR professionals identify which employees are at risk of leaving the company and provide personalised recommendations to improve employee engagement.
- Talent development and training: AI can help identify skill gaps in the workforce and provide personalised training recommendations to employees.
- Workforce planning: AI can help HR professionals anticipate future workforce needs and plan accordingly .
- HR chatbots and virtual assistants: AI-powered chatbots and virtual assistants can help HR professionals answer common employee questions, provide support for routine tasks, and more.
Use cases: HR based AI use cases enabled by text-based processing of materials related to human activities such as personal performance assessments, as well as details of individual involvement across the operational and business realm. The increasing ability of language-aware AI to extract and evaluate these materials highlights use cases where AI provides agility, scale, and consistency while unlocking efficiencies and savings.
Career management – Analysing data to align individuals against career and role profiles to support employer and management reviews and feedback.
Recruitment and hiring, processing applications and CV’s – Processing resumés and applications to stream and categorise these and reduce administrative overhead
Self-service applications – Chatbot based interviewing and assessment, prompted by candidate responses and resumés.
Employee Performance analysis – processing employee and management performance review assessments, highlighting and scheduling key-actions, making policy-based recommendations
Email and messaging generation – Generation of personalised email communications for HR purposes
Social Listening – reviewing social media for employee interaction and discussion of company related matters, identifying preferences and issues.
Job description generations – creating job descriptions for advertising campaigns
ASK HR Self-service – HR related chatbots managing employee queries and requests
Retention management – examining human capital marketplace conditions to react and adjust to changes in expectation for employee roles in order to protect retention levels.
Identification of at-risk employees. Identifying factors in performance review, attendance records, social media, etc indicating at-risk employees for well-being, retention, or grievance purposes.
Voice of Employer processing and surveys – analysis of third-party VOE sites, online news, social media, etc to determine employee satisfaction levels or trends
Self-service on-boarding – Chatbot based onboarding, indoctrination sessions and intake administration interacting with candidate responses.
Current state AI for customer service
- Chatbots – Current approaches using scripted chatbots in rule-driven scripted engagements.
- NLP and Sentiment analysis – Current approaches also use existing AI based approaches driven by NLP, sentiment analysis and keyword recognition.
AI trends for customer service and their business impact
- Customer Engagement – AI customer service for higher customer engagement: AI helps organisations engage customers more effectively, increasing their loyalty, value, and retention. AI enables personalized service, and can evaluate customer lifetime value estimation,
AI trends for HR and their business impact
-
User experience process improv ements – AI solutions for user experience, and process improvement.
AI can help obtain insights and guide decisions, ensuring optimal user experience through chatbots and sentiment analysis, improve processes through intelligent document processing, workforce management, post call wrap-up, and robotic process automation. - Next generation AI-powered chatbots – For more effective self-service AI driven Chatbots are capable of handling complex customer queries and interactions. These chatbots also integrate with other channels and systems to provide seamless and consistent service while reducing costs, increasing satisfaction, generating leads, etc..
- Automation – Customer service automation for faster ROI for SMBs: Small and medium-sized businesses (SMBs) can benefit from customer service automation as it can help them simplify their tech stack, empower their agents, and adopt new technologies faster. Customer service automation can also help SMBs scale their operations and compete with larger players.
Use cases: Customer service use-cases using AI are focused on streamlining the customer service experience and engaging with the customer for all their needs, anticipating and responding to their preferences and engaging with them in a personalised and highly aware manner.
Call Classification and routing – intelligent responsive call handling
Speed and Accuracy of information – Surfacing customer information and answers quickly via predictive methods
Social Media Listening – Correlation and usage of customer details from social media
Processing communications, surveys, and reviews – Using advanced language to extract specific and general customer viewpoints and concerns from textual information.
Voice recognition and authentication – AI driven voice-based sophistication for voice recognition and authentication
Email generation and follow up activities – Post contact activities such as email generation, customer write-ups and scheduling of proactive customer care steps.
Proactive engagement – Increased customer satisfaction through proactive customer services engagement, initiating actions, call-backs, etc.
Self-service customer solutions – Use of natural language features to permit naturalistic customer self-service.
Agent Efficiency – Agent efficiency through streamlined and automated customer engagement handling.
Recommender systems – Support customer care and revenue generation with recommender systems.
Intent discovery – More accurate understanding of customer contact point motivations.
Customer support summaries and summary frequency – collating multiple data sources for customer support and CRM summaries. Identifying events pertinent to specific customers.
Fully automated administration routines – Streamlined customer interaction via language enabled fully automated administration routines
Tuning best practice answers and consistency – Predictively determining best practice answers and responses, enabling customer service consistency.
Identification of at-risk customers – Preventing churn by identifying at-risk and dissatisfied customers.
Current state AI for data management
- Data management is the process of collecting, storing, organizing, and using data effectively and efficiently. AI already helps with data management by automating and simplifying tasks such as data discovery, integration, cleansing, governance, and mastering. AI is also improving data understanding through existing NLP and sentiment analysis, as well as enhancing data security and privacy.
Data mastery
- Emerging discipline – Data Mastery is a term used to describe the ability of organisations to leverage data to gain a competitive advantage. According to research only 16% of organisations currently have the tools, technologies, people, processes, skills, and culture to successfully derive value from growing volumes of data.
- Scope : Wider and Deeper – With AI the possibilities for data mastery have greatly expanded to include the myriad interconnections and increased subtlety accessible to AI methods.
- Amplify maturity – There are many emerging approaches to Data mastery available to help organisations improve their data maturity and evolve their data maturity model to better manage data resources, AI methods are fundamental to that process.
AI trends for data management
- Data Strategy – AI is already central to data analytics, the process of analysing data to gain insights and support decision making. AI driven data management can evolve the basis for analytics, assisting strategy, by discovering and highlighting data from multiple sources for evolving insights.
- Data Integration – AI can drive data integration, combining data from different sources and formats into unified views. AI uses techniques such as entity resolution and record linkage to match and merge data from different sources.
- Data Quality – AI can also enforce data quality, ensuring that data is accurate, complete, consistent, and reliable. AI uses techniques like anomaly detection and data cleansing to identify and correct data errors.
- Data Cataloguing – Data cataloguing, the process of creating and maintaining a searchable inventory of data assets. AI can use techniques such as natural language understanding and semantic analysis to extract metadata from data sources, such as descriptions, keywords, tags, and categories to identify data relationships of business value.
- Data Security – AI can also monitor data access and visibility for anomalies and potential threats, and enforce data protection policies, augmenting many cybersecurity controls and protocols.
Use cases: AI has enabled an enhanced level of data mastery and depth for data management. Use cases here are delivering improved data quality and enabling improved depth/finesse in unlocking business benefits from data as Ai has enabled the sophisticated detection of business events from data.
Data manipulation and alignment for quality – AI can search for and align data to apply data quality checks and adjustments.
Merging data from disparate sources – AI can merge data from differing sources, such as legacy silos to give overarching visibility, or to enrich data held in critical systems (e.g.,. ERP, KPI, CRM, etc)
Automated indexing and cataloguing of data – AI can review data, creating indexes for business specified purposes as well as cataloguing data to highlight features of business interest
Advanced data queries and filtering – AI can be trained and tuned to hold business data and enable sophisticated overarching data queries and data extraction, such as filtering.
What-If scenario processing – AI can establish data models representing differing business What-If scenarios. These differing viewpoints enable informed strategic evaluation of alternatives
Data obfuscation – AI can detect and mask sensitive or regulated data, obfuscating these for training or application testing purposes.
Optimising data access costs and data relocation strategies – Accessing and moving large amounts of data can incur significant cost with cloud-based infrastructure. AI can calculate optimal storage strategy and lower costs based on anticipated usage, relocating data to lower cost storage tiers as needed.
Enhanced data security – AI can apply knowledge of up-to-the-minute security threats and evaluate these against the as-is configuration to anticipate and prevent cyber attacks
Synthetic data creation – AI can model existing real-world datasets and generate synthetic data for application testing, or for business stress testing.
Current state of AI for financial control
- Data analytics – Data analytics KPI’s are often focused on financial control as well as other operational considerations embodied in KPI’s. Specific current AI approaches are focused on anomaly detection detecting variance from limits and controls set for conducting business.
- Decision Support – Current decision support approaches use model-based approaches to detect actionable patterns using AI Machine learning approaches.
AI trends for financial control and their business impact :
- Demand / Revenue Forecasting – Using both internal and external sources of data, models predict demand and associated revenue across a variety of dimensions including business unit, product line, SKU, customer type, and region.
- Anomaly and Error Detection – Anomaly detection uses a series of machine learning (ML) models to highlight transactions or balances that are in error or potentially violate accounting principles or policies. A comprehensive solution will also include real-time analysis during data entry preventing errors from entering the workflow and avoiding costly downstream corrections.
- Decision Support: AI predictions based on hypothetical business scenarios are used to evaluate potential outcomes and decision strategies.
- POC Revenue Forecasting: AI models forecast the percentage-of-completion metrics (e.g., hours, cost, units, weight, etc.) to predict POC revenue and the total completion effort remaining.
- Cash Collections: AI can forecast at-risk customer payment performance and initiate proactive collection actions.
Use cases: AI is moving financial control to an ever-increasingly real-time financial view of a business to support high-fidelity decision making. AI can more fluidly combine and extract information from multiple sources providing insights in response to changing business conditions.
Demand and revenue forecasting – AI can predict demand and revenue based on marketplace analysis, anticipating costs against budgets and cashflows.
Anomaly and error detection – marketplace and business conditions can be examined by AI to detect or predict unusual patterns and anomalies.
Invoices and paperwork – AI can process business paperwork, both generating outward items (invoices, purchase orders, supplier agreements, etc.) as well as processing received items (supplier invoices, correspondence, agreements, etc.)
Decision support and what-if analysis – AI can support what-if scenarios to allow evaluation of differing financial strategies to support strategic decision making.
Point of cycle revenue and accounting – Increased data agility and handling abilities allow much improved point-of-cycle reporting to better respond to business events.
Collections activities / creditor analysis at-risk accounts – AI can model creditor behaviour to highlight or anticipate at-risk accounts.
Fraud detection and preventions – AI can monitor internal financial processes and patterns to detect and avoid fraudulent activity.
Trading decision support – AI can review trading activity to detect errors ensuring that trading processes are being correctly executed.
Trends and triggers – AI driven data management can detect important business events and trends to trigger appropriate responses, such as critical price increases.
Risk Management – AI can be applied to the minutiae of a business risk register and review available data to gather relevant detail for review.
Credit Scoring/Customer intelligence – AI can utilise external data sources and market intelligence to support decision making such as credit scoring
Current state AI for software development
- AI is within the contemporary SDLC – Currently, meaning until 2023, software development has delivered AI as a solution centred technology within established best practice SDLC (Software Development Life Cycle) and delivered according to established methodologies and paradigms. Current AI methods utilise AI approaches such as machine learning, natural language processing, and sentiment analysis. Delivering solutions using these methods and approaches is achieved in the same manner as the delivery of solutions based on differing, non AI-based, methods and techniques (notwithstanding additional SDLC based activities, required for AI methods, such as model training, etc.)
- Development-maintenance cycle – Typically software development projects are relatively short periods of development followed by much longer periods of ongoing maintenance and modification until circumstances force the obsolescence of the application. The ROI for software development requires sufficiently usability over this longer period to offset the costs of software development. Typically, during this period maintenance can become increasingly difficult as complexity compounds within the application.
AI trends for software development
- AI altered SDLC – The advent of LLM based AI will significantly alter the contemporary SDLC. It is already evident that development coding, testing, and software integration will be significantly accelerated by the usage of LLM-based AI. Collectively these activities represent a very significant proportion of the SDLC in terms of cost, duration, and risk. A shift toward consistent approaches in the modified SDLC will result in modified processes, reduced costs and better managed complexities.
- Differing development/maintenance ROI dynamics – The ability to generate applications more rapidly will allow the development/maintenance cycle to be changed. ROI will be changed as maintenance will benefit from the same SDLC improvements as development, thus applications will have a longer in-service lifetime, or can more easily be replaced.
- Skills shortages – Frequently relevant technical skills are in short supply, the AI supported SDLC will greatly reduce the impact of these shortages.
- Impact on COTS – (Commercial off-the-shelf Software) – The altered SDLC will allow more competition in the COTS market as the barrier for entry for solutions will be lowered. This is likely to result in more competitive software licensing in the future for all software.
Use cases: For technology-based activities, such as software development, AI use cases represent an increased awareness of practice activities at all levels, allowing a more tool-based and consistent approach for these activities. Well-ordered and structured activities will be achievable to a high level of precision with very significant improvements in productivity and overall risk reduction.
Automating code review, bug fixing, testing, and compilation – AI can examine Existing or legacy code to detect bugs, make amendments, generate test data, or migrate to a different coding language
Generating code automatically – LLM based code generators will be able to generate draft software in a consistent and instantaneous manner.
Creating better SDLC toolsets & frameworks – AI supported software engineering toolsets and frameworks will bring benefits and improved ways-of-working.
Enhanced User Experience – The increased ease of creating delivered software will allow more A-B testing to take place, yielding higher levels of User-Experience.
SDLC decision making – AI, by rapidly centralising salient issues will Improve SDLC day-to-day decision-making, error management, and task time estimation.
SDLC design activities – AI is also changing the way software engineers think and work, pivoting from design to platform thinking, focusing on how platforms function in goal-oriented design.
Automated interface negotiation – AI can generate interfacing code between applications by examining their interfaces or wrap existing code in APIs to conform to standard interfacing patterns.
Documentation generation – AI can generate documentation for program code for the use of support staff, end-users or customers.
Generate SDLC deliverables – AI can draft or check user stories, acceptance criteria, and business requirements, while conforming to all applicable standards, conventions and guidelines.
Knowledge and courseware generation – AI can generate detailed engineering specifications or educational courseware from application code.
Cost optimisation – Understanding production runtime infrastructure at design time AI can estimate production runtime costs to better inform application design
N.B. – Some of these use-cases are things that will change because of AI impact, such as the enhanced toolsets, improved decision making and design thinking etc.
