The short version

Waste in insurance has a precise definition: valueless complexity.

Insurance is a complex business, and much of that complexity is the product. Clients pay for bespoke cover, careful wordings and judgement on difficult risks. But around that sits a second kind: rekeying, reconciliation, duplicate checks and processes that differ for no reason anyone can remember. Nobody pays for it, and nobody would miss it.

Lean thinking gives a way to tell the two apart and remove the second. It matters more now than ever, because AI applied to valueless complexity does not remove it. It makes it faster, and harder to see.

Two kinds of complexity

The test is whether the client, or the capital behind the risk, would pay for it.

Valuable complexity Valueless complexity
What it is Complexity in the risk and the cover Complexity in how the work gets done
Examples A bespoke wording, a layered programme, a judgement call on an unusual risk The same data keyed into three systems, a reconciliation between two of them, a sign-off nobody reads
Who values it The client, the broker, the capital provider Nobody
What to do with it Protect it and give it more time Remove it

The specialty market is right to defend the first kind. The mistake is using it as cover for the second. "We are a complex business" is true of the risks, and it is used to excuse the processes.

The same from outside, many ways inside

From the outside, every class of business in an insurer does the same things. It accepts or declines a submission, provides a quote, issues a policy and renews it. It collects premium, and cedes premium out to reinsurers. It takes first notice of a loss, settles the claim and collects the reinsurance recovery. At the end, it reports its financial results.

Inside, there are often numerous ways of doing each of those, across business units and within them.

The cause is the way our industry is siloed by class of business. The producing broker holds the client relationship. From there, the client's risks are split by class and passed to placement brokers, then to carriers, then to reinsurers. This is especially true of large, complex risks. It is less pronounced for smaller clients, where a packaged policy is enough.

The split runs deep. Within one insurer's marine division, hull and war underwriting can work very differently from the cargo team. In professional indemnity, the process for architects and engineers can be quite unlike the process for software integrators. The underwriters are not aligned across those classes, so each has built its own way of working.

Some of those differences reflect the risk. Many do not. The client, the broker and the reinsurer see the same quote, the same policy and the same settlement, whichever team produced them. A difference that nobody outside the firm can see is complexity the firm pays for alone.

Where valueless complexity hides

Lean thinking grew out of the way Toyota builds cars. It describes eight kinds of waste. Adapted to insurance, and with batching added, they look like this.

Waste What it looks like in insurance
Waiting A submission sits in an inbox, or a claim waits for a coverage check
Handoffs Work passes between underwriter, assistant, operations and finance, with a queue at each step
Rework and defects Queries back to the broker, corrected bordereaux, endorsements issued to fix errors
Over-processing The same risk checked several times, and referrals for decisions well within authority
Duplication Data rekeyed from email to rating tool to policy system to spreadsheet
Inventory Backlogs of unbooked premium, unmatched cash and unsigned documents
Batching Work held for the monthly run, the weekly committee or the full file, when it could move today
Needless variation A different process for each class, office or team, with no difference in the risk to justify it
Unused talent Skilled underwriters and claims handlers spending their day on administration

The last is the most expensive. The market's scarcest resource is experienced judgement, and valueless complexity spends it on work that needs none.

Why it survives

Valueless complexity is rarely designed. It accumulates, and three things keep it in place.

  • Every exception once had a reason. A control added after an error, a spreadsheet built to bridge two systems, a process inherited with an acquisition. The reason goes and the step stays.
  • Good people absorb it. Experienced staff work round the problem so well that it never reaches a board paper.
  • Nobody owns the whole flow. Each team improves its own step. The waste sits in the gaps between them, where no one is accountable.

Lean before AI

The order matters: fix the flow first, then automate what remains.

  1. Map the value stream. Follow one piece of work from start to finish, and record every step, handoff and wait.
  2. Ask of each step: would the client pay for this? If not, it is a candidate for removal, however long it has been there.
  3. Remove before you automate. Take out the step, the duplicate entry or the extra check. This often needs a decision, not a system.
  4. Standardise what is left. One way of doing the work, unless the risk itself demands a difference.
  5. Then apply technology and AI. Now automation speeds up work that is worth doing.

Skip the first four and AI gives you the same waste, faster. A model that reads a submission in seconds saves nothing if the data is still rekeyed three times afterwards. Worse, automation hides the waste inside the system, where it is harder to find and more expensive to remove.

Who removes it

Not a transformation team. The typical insurer has perhaps 5% of its people trying to change the other 95%, and it rarely works. The people who hold the complexity are the ones best placed to see which parts deliver value, and to let the rest go.

That is a change in who works on the business, and it has its own piece: The 5% and the 95%.

Questions for the executive team

  • Which of our processes would a client refuse to pay for, if they saw them?
  • Where do two of our teams do the same job in different ways, with the same result for the client?
  • How much of our underwriters' and claims handlers' time goes on work that needs their judgement?
  • Who owns our most important value stream from end to end?
  • When did we last remove a control, a report or a step, and not only add one?
  • Are we about to automate a process we should first simplify?

How Bind helps

We help firms find their valueless complexity, remove it and keep it out.

  • Operating model: value stream ownership, and teams able to find and remove waste in their own work.
  • Digital platform and architecture: platforms simplified to match the simplified flow.
  • AI readiness and adoption: AI applied to work that is worth doing.