Where to invest in the age of AI

We are asked a lot by our customers and partners what we should be doing with AI? Can you help us craft our AI strategy? It seems the right time to write down some of these thoughts on paper, and articulate the position of B1, putting our thoughts into one document to help others on how to invest, plan and help support growth of their business in the world of AI.

The key should be to set up the enterprise now to get better, as AI gets smarter and unlocks more power.

Why now

If you don’t have a strategy that focuses on leveraging AI, you need one. At this stage it’s clear that the intelligence AI brings, opens up new flexibility to the workflows of future business.

Every business talks about re-invention, this should not be your focus. Don’t begin by assuming AI requires business re-invention. Start by focusing on the fact it is revolutionary in the sense that it changes the economics and process of how you think about your approach to the workflows that back your business. What follows is a structured approach to invest in business understanding, followed by rules, data, tools and workflow to structure your future approach to supporting the business.

Why Orchestration & Context is Key Business is all about connection and orchestration. Taking orders. Approving purchase. Planning demand. Fulfilling customer orders. Closing sales. Onboarding employees and customers. None of this changes.

What changes is the procedural nature of the past. We focused on a defined process that required our intervention at each stage. We built bespoke automation processes to aim to streamline, or ease the flow. We built custom systems, custom integrations and custom approaches to respond to the demands of the business. We didn’t always understand why things worked the way they did, or how they were actually working.

Investment Area #1 - Invest in Making the Business Accessible The key to unlocking the future of agents and AI is to invest in key system connections. Without a connected system, it is impossible to deploy Agents, Agentic Frameworks and automations that can orchestrate flows in an automated manner.

This should be focus #1. A first step is to undertake an AI assisted mapping exercise to develop a coordinated map of your systems, workflows and connections to identify the gaps where information, business process endpoints and logic is not exposed properly, or buried inside a manual process.

Key Outputs:

1. True as built system architectures
2. As built business capabilities
3. Overall business mappings and workflows for existing current state processes
4. Business capabilities / tool catalogue

Investment Area #2 - Change the Economics of Solution Delivery Teams can move faster to develop solutions than ever before. To do that, they require the input from the initial mapping. CTO’s should be working with teams around AI assisted delivery flows so they understand how they can leverage AI to get the existing work they do done faster.

Armed with the mapping, business rules, workflows and system connections, it unlocked a second level of delivery for the business. Teams can now move faster with confidence in the current state, delivering more solutions faster to key existing workflows.

The cost of building out solution ideas is decreasing continually CTO’s should be pushing teams to be closer to the business problems they are solving. They should be working with clearer intent on what they are trying to achieve. So instead of setting up around the IT delivery timeline, they should be setting teams up to understand the approach, intent, context and constraints better. It’s not enough to just give Co-pilot or Claude to a team, we need to have them re-think how they get the context that matters. It’s now faster then ever to generate lines of code, but still just as hard to define what they should be doing.

Key Outputs:

1. AI Engineering tooling: Standards and repositories for agent usage.
2. Change the work intake process: Better definitions of intent, requirements and processes. Teach the business how agents will be executing in an assisted matter with engineers, so they understand better how to describe what they need
3. Automation Testing and Validation Harness: A defined approach to validation and automation to ensure that execution is confirmed.
4. Engineering approach focused on intent, orchestration and validation.

Investment Area #3 - Agentic processes as part of the business

While investments in #1 and #2 are maturing. The CTO and organization should be planning for a future where some processes move from manual procedural workflows to semi autonomous agentic processes. This is not as fancy as it sounds. The word agentic seems to imply magic, but armed with the components of investments in business rules, workflows and system connections, we can start to invest in semi autonomous workflows that unlock human intervention at every stage.

Is this really different? The reality is this roadmap is no different then what we should have been doing since the beginning of time. The unlock today is AI has the capability to help us understand complex systems, map workflows effectively so we can then truly bring probabilistic steps to scenarios.

Why is probabilistic a benefit? The exciting part about moving from a procedural process where we defined all actions explicitly is that Agents benefit from intelligence. Armed with the investments in #1 and #2 it gives Agents enough context, system connections and abilities, access to existing systems that they can actually do work.

Where the magic happens. At each stage, structuring your investment in this order will yield compounding results both on solution delivery, overall team understanding, and aha moments that will unlock moments like: “why are we doing it like this today?”. This process is what the best in the business are doing today at Anthropic, Open AI and other Forward Deployed Engineering firms. But you don’t have to wait for their sales team to sell you on a Forward Deployed Engineer, you can start to bring these processes in house today and set the vision for your business, helping them understand that they can do more, with less manual workarounds and deliver a better result to your customers.

The important shift is from explicitly programmed workflow logic toward systems that can decide what to do next based on an objective, context, constraints and available tools.

Key Outputs

1. Agent Use Case Portfolio: A list of potential use cases that agents could leverage judgment to automate.
2. Defined Agent Harness and Frameworks: an agentic framework that provides the foundation capabilities for Actions, Tools access, Workflow orchestration and policy invocation and observability.
3. Autonomy Roadmap: A roadmap for the autonomy roadmap from Human led today, AI assisted in future, to bounded autonomy.

The goal of this roadmap is to structure your investment in AI technology to take advantage of the improvements that will come in time. As intelligence improves with these foundational blocks in place your organization will be able to take advantage of them at each improvement level. By focusing on business workings, systems, rules and data as agent and AI capabilities improve you will be able to continue to structure your investments to take advantage of the increasing capabilities that will be provided.

If investment is made in this order, we are confident that enterprises will be ready for tomorrow’s models, to make organizations faster, more capable and more autonomous without having to re-start.