Observations
The myriad of announcements each week regarding embedded AI and the integration of agentic services by business platform vendors align with IBRS’s predictions that the main use of AI will be embedded within core business solutions, integrated directly into established operational workflows, rather than deployed as an external application layer.
Embedding AI into core business applications is not a fad. It is not the vendors just jumping on the AI bandwagon. It is the natural location for all forms of AI. IBRS strongly advocates that AI be viewed as a software capability – a component – rather than a solution or product in its own right. When viewed this way, business executives and business analysts can get past the hype and begin mapping where different AI capabilities can be applied to business processes. In the majority of cases, the processes that can be improved or streamlined sit within existing core business platforms.
What Does This Mean For Your AI Strategy?
The AI that will generate the bulk of productivity gains is likely already being developed by business solution partners. Investing in complex, custom AI platforms and bespoke AI projects may only be buying into new technical debt that, within the next three years, will be entirely superseded by core business platform investments the organisation has already made.
In addition, the more established core business solutions will likely provide better and more cost-effective AI than building customer solutions for several reasons:
- Fit-For-Purpose: Leading business platform vendors can tap their deep understanding of data types and patterns, so they can refine or even train smaller, highly specialised generative and machine learning (ML) models. Smaller, fit-for-purpose models outperform larger models in terms of run costs, speed, and reduced hallucination rates. Platform vendors will introduce many small, highly specialised models, each addressing specific business needs. Training to refine bespoke models within an enterprise is an expensive and high-risk proposition.
- Prioritisation of Economic Returns: Business platform vendors will prioritise the processes where AI services will see the most significant utilisation, in no small part because they want to maximise consumption of such services – this is how they will see returns on their AI investments. The result is that while many business platform AI services will be less exciting – or even invisible – to business stakeholders, they will provide continual, frequently run behind-the-scenes efficiency gains. This contrasts with the current situation, in which AI development teams struggle to demonstrate business value for AI initiatives.
- Contextual Data Gravity: Business platform vendors are in a unique position: they define the data structures and database architecture. This means they can leverage advances in Cloud databases that place vectors (the way GenAI encodes information) and graphs (contextual links between data) on top of their entire applications’ relational and transactional data. In essence, the vendors are uniquely positioned to AI-enable their entire stack, turning every data element into a native input for AI services. This goes well beyond simply fetching data from the platform via an API call and passing the results into an AI service. It will enable enterprise data to be part of a business platform-specific AI model. Currently, the only way for an enterprise to achieve a similar capability is to extract all platform data into a data warehouse and manually build AI pipelines – a costly, complex challenge.
- Development Velocity: Not only do many platform vendors have a larger pool of software development talent and deliver new capabilities at a regular cadence, but they also benefit from being able to leverage AI based on their internal knowledge and software architecture libraries that are simply not available to clients. This means most – though not all – platform vendors will deliver AI-enabled processes more quickly than their clients can deliver similar capabilities. This does not mean they will prioritise the services any one particular client desires, but the cadence of new AI capabilities will be faster overall.
AI Becomes Invisible
The vendors’ inclusion of AI into their platforms’ business processes will render a significant portion of AI capabilities effectively invisible to many staff. While the process may become more streamlined, much of the AI (ML, generative and graph lookups) will be working in the background, and only presenting the results, summarisations, and recommendations near the end of a series of automated actions. Humans will still be in the loop… but right at the end of the loop. Prompting and manual reworking of AI service outputs will decline sharply for many staff as this trend towards invisible AI continues.
Infographic: Examples of the Trend Towards Invisible AI
Following is a sample of key announcements from major business solution vendors that clearly shows this trend in action. The announcements are accelerating. (See downloadable infographic below)
| Date | Vendor and Product | Summary of AI Announcement |
| May 2023 | ServiceNow | Now Platform: Unveiled Now Assist, using GenAI to summarise tickets and streamline employee/customer requests. |
| June 2023 | Adobe | Adobe Experience Cloud: Embedded Sensei GenAI to generate and tailor text for marketing campaigns across channels. |
| July 2023 | HubSpot | HubSpot AI: Introduced AI assistants to instantly draft marketing emails, social media copy, and blog posts. |
| September 2023 | SAP | SAP Joule Copilot: Embedded a natural language assistant across its Cloud portfolio to automate business tasks. |
| September 2023 | Workday | Workday HCM: AI helps managers write performance reviews and create personalised employee development plans. |
| September 2023 | Intuit | Intuit Assist (QuickBooks): AI provides cash flow insights and automates invoice reminders for small businesses. |
| October 2023 | NetSuite (Oracle) | NetSuite Analytics Warehouse: AI predicts inventory demand, optimising stock levels and reducing carrying costs for supply chains. |
| October 2023 | UKG | UKG Pro Suite: AI helps managers with intelligent shift scheduling to align staffing with business demand. |
| April 2024 | Microsoft | Dynamics 365 Sales: Copilot provides sellers with timely customer insights to help expedite their sales deals. |
| April 2024 | Microsoft | Dynamics 365 Finance: AI-guided rules ease financial dimension setup and increase automation in bank reconciliation. |
| May 2024 | Zoho | Zoho CRM (Zia): AI suggests the best time to contact leads and predicts deal conversion probability. |
| June 2024 | Ceridian | Dayforce Co-Pilot: Helps managers create compliance-aware job descriptions and summarise complex HR reports. |
| September 2024 | Salesforce | Einstein Copilot: Deploys conversational AI agents to automate sales updates and create service knowledge articles. |
| October 2024 | Infor | Infor Enterprise Automation: AI automates accounts payable invoice processing, from ingestion and extraction to validation. |
| November 2024 | Sage | Sage Intacct: AI-powered anomaly detection flags unusual transactions for review within the general ledger. |
| January 2025 | Unit4 | Unit4 ERPx: An AI assistant helps with project resource planning and automates expense report creation. |
| March 2025 | Coupa | Coupa BSM Platform: AI analyses business spend data to identify savings opportunities and supplier negotiation points. |
| June 2025 | Zendesk | Zendesk Suite: AI automatically triages support tickets and suggests macro responses to speed up resolutions. |
| August 2025 | Freshworks | Freshworks Freddy AI: Freddy Copilot summarises customer conversations to streamline agent handovers and improve service continuity. |
Next Steps
- Engage with your core business platform vendors to fully understand their AI and agentic services roadmap:
- How they are currently embedding AI services, and which models they are leveraging.
- Their near-term plans for agentic AI services, and especially low-code tools for development agents within their platform.
- Their longer-term architectural roadmap, down to the database layer.
- How they intend to charge for the consumption of AI services.
- Assess the security and compliance frameworks supporting these AI capability roadmaps, particularly concerning data privacy and ethical AI use.
- Request detailed demonstrations from vendors, focusing on how agentic AI integrates with existing workflows and data structures.
- Working with line-of-business stakeholders, evaluate the specific use cases of embedded AI within your organisation’s current HCM, ERP, and CRM solutions to identify potential operational enhancements.
- Map the vendors’ roadmaps to your organisation’s broader organisational strategy and ICT strategy. Align the vendor roadmaps with long-term organisational digital transformation strategies. Then determine how these fit within your organisation’s AI strategy (if it is separate from the ICT strategy).
Infographic: Examples of the Trend Towards Invisible AI
