AI in all its forms (analytical, generative, agentic, et al) promises to redefine how work gets done. It will make recommendations, support decisions, generate insights, execute workflows, and learn from that execution experience.
But there’s a catch – AI is only as powerful as the data behind it, and this is true for all business process automations and augmentations.
We continually build incremental systems, workflows and SOPs to manage the exceptions originated by data integrity problems. We are creating our own ‘failure demand’ activity that is ultimately wasteful and unproductive.
This is exacerbated by our aspirations for, and adoption of, AI. We need to go back to the source, the data. Because the mantras of ‘data centric’ and ‘data-driven’ are often presented as truisms, without depth. Data is a core business issue, but it often seems amorphous, ambiguous, and enormous.
Front of mind in the C-Suite
To focus on maximum business impact, four key, current, tectonic shifts are getting to the front of mind in the C-Suite – and will cause an earthquake if we don’t address them:
- Geopolitical shifts are causing companies around the world to realise they cannot trust their data on country of origin, tariff codes and material categories. The data lacks acceptance, accuracy and integrity. In addition, the mechanisms for creating and maintaining clean data are ambiguous or failing.
- Efforts in business transformations and end-to-end process/value stream integrations are failing due to silos of inaccurate, conflicting, and duplicated data.
- Global ERP standardisation initiatives – driving multiple regional or BU specific ERPs to a single instance of S/4HANA, for example – are highlighting the critical need to standardise the data that drives the business. It can no longer reflect purely local or functional requirements.
- And, of course, the hotly anticipated benefits of AI initiatives are failing to appear due to ambiguous, erroneous, redundant and duplicated data.
Not just a technology issue
This data problem is often put on the too-hard list, or just regarded as a technology issue, or because previous attempts have resulted in lengthy initiatives that, quite frankly, feel like trying to boil the ocean. But, it doesn’t have to be that way.
The key priority is the essential ‘Master Data’, sometimes called Reference Data. This drives our business processes and fuels the machinery of our automation and augmentation initiatives, such as:
- Customers
- Suppliers/Vendors
- Products
- Materials
- People and Assets
- Locations
Master data reliability
There are typically one or two of these master data domains that are directly relevant to each end-to-end process – customer to cash, source to pay, plan to make, hire to retire. It makes sense to align initiatives with these process leaders for context, efficiency and buy-in. However, typical master data reliability issues fall into a number of areas, as our own work with customers evidences:
Duplication – deliberate or planned (yes, there can be bona fide reasons for duplication), it impacts customer experience, asset utilisation, process efficiency, errors and re-work, leakage of revenue, cash, cost and balance sheet integrity. Often, duplication is not immediately apparent but implied by common data. It is not unusual to see up to 20% duplication in active domains.
Quality and Integrity – completeness and accuracy of master data is critical to essential procedures such as tax/VAT IDs, bank details, agreed payment terms, country of origin, industry, export licencing, sanction checks, etc. Up to 50% of master data is not complete and/or accurate enough to enable effective process operations.
Inactivity – key business partners, people, places and things with no activity over a defined historical period, indicating poor asset utilisation or candidates for inactivation/archival. Up to 85% of inactive customers and suppliers still have active master data that anyone can create activity for, erroneously, because they have been left unblocked.
Activity – master data does not exist in isolation – it is amplified by the transactional events that are related to it. So, activity over a period accumulates accurate or inaccurate data on customer and product experience, assets, liabilities, P&L, cash and working capital.
This is why master data is at the core of productivity, because all the issues identified create errors, re-work, queries, escalations, credits, reversals, etc – all those things that absorb between 25% and 50% of total effort in GBS, Shared Services and Finance Operations.
Data foundation fit for purpose
Be aware that at least one of the four key, current, tectonic shifts above is affecting every business, maybe more, so we cannot afford to keep kicking the can down the road. We need to start talking about Master Data NOW! We need to take action and work on it – and assess is the critical data foundation fit for purpose?
We need to get clean and stay clean – only then can we be data ready, at speed, for AI-enabled transformation. Only then will our revamped end-to-end processes deliver on their promise of enhanced business value and P&L impact.
I have a favourite image (as above) that encompasses these business realities, a ripe apple in front of a mirror. The reflection shows a beautiful rosy, ready to eat to apple. The apple itself, in the foreground, has a large bite out of it, the flesh slightly oxidised and less appealing. But the mirror only reflects what is in front of it and the same is true of the deception and indeed illusion that more data is automatically a good thing – because data can only be partial, biased and retrospective.
Consider Solutions focus on delivering business performance outcome improvements, not just reports, that bring tangible P&L, Working Capital, Balance Sheet and Risk Management impact – through a proprietary end-to-end (E2E) business process approach and smarter use of technology in the context of human behaviour, data insights and decision support. The ‘Pareto Principle’ is a key to success in all of these initiatives; the 80/20 rule by which 80% or more of desired outcomes can be delivered with 20% or less of the cost, effort and cycle time.
